Is Open Source too Complex?

Jason Pillai writes to tell us ZDNet is reporting that at last month's Microsoft Worldwide Parter Conference in Boston Ryan Gavin, director of platform strategy, claimed that one of the big downsides to open source is complexity. From the article: "Gavin noted that the flexibility of open-source software in meeting specific business needs also means systems integrators and ISVs have to grapple with complexity costs. 'It's challenging for partners to build competencies to support Linux, because you never quite know what you're going to be supporting,' he added. 'Customers who run Linux could be operating in Red Hat, [Novell's] Suse, or even customized Debian environments,' he explained. 'You don't get that repeatable [development] process to build your business over time.'" More than once I have had complaints that my setup is more difficult than necessary. Is open source really that much harder, or just different than what most are used to?

Comments

All Software is complex.

[Whiney Mac Fanboy]

Article should read:

"Gavin noted that the flexibility of proprietary software in meeting specific business needs also means systems integrators and ISVs have to grapple with complexity costs. 'It's challenging for partners to build competencies to support Windows, because you never quite know what you're going to be supporting,' he added. 'Customers who run Windows could be operating in 98, XP SP1, SP2, Vista or even customized 2003 server environments with god alone knowing what browser version they're using' he explained. 'You don't get that repeatable [development] process to build your business over time.'"

Because, lets face it - what Gavin is saying here is that proprietary software vendors find it hard to develop for linux. *shrugs* Maybe, that is not the same as saying that developing for Open Source is complex.

MS - time to face it, almost noone apart from you is making alot of money selling proprietary sofware (alone). The real cash is in services, services, services.

Re:All Software is complex.

[khakipuce]

Or put another way, a guy from Microsoft, who has probably never configured or operated any of the systems he mentions, is telling a group of people, who also have probably never used those systems, that it's really scarey if you move away from Microsoft...

And this is NEWS?

For my sins I have used a lost of operating systems over the years and they all have their pros and cons, the one thing that seems common across them is that the more scarey they look the less likely they are to break because people don't mess with the difficult ones. Most failures are caused by human error (it's just that no one admits to it) and making server OS's look familiar tempts people to fiddle.

Re:All Software is complex.

[Anonymous Coward]

Because, lets face it - what Gavin is saying here is that proprietary software vendors find it hard to develop for linux.

No, he's saying that thorough QA and support to the standard that enterprise customers want is more work the more platforms your application has to run on. It's too much risk to say "we support all Linux".

The solution, as the article hints, is to say "we support our app on the latest RHEL only". And then you're back in a low risk situation.

Re:All Software is complex.

[peragrin]

ah but that's the good thing about Linux that you can not say about MSFT.

You as a vendor only supports RHEL, but that means your customers can get their Linux from Red Hat, CentOS, White Hat, or any of the other firms that take RHEL sources remove the trademarks and redistribute the binaries.

Your still not locked to anyone vendor for support or services.

Re:All Software is complex.

[Enderandrew]

I think you homed in on what the real message is here. And there is some truth to it.

For a large software developer, they want to reach a wide audience. That is why they develop for Windows, which has the largest user base. Even if they want to reach out to a Linux user base, there are so many different distros and ways of doing things, that you cut that audience into even smaller shares.

I'm all for freedom, but I find it silly that different distros keep configuration files in different locations, use different init scripts, use different install methods, have varying level of compliance with the LSB, are focused on either Gnome or KDE predominately, etc.

It does make it more difficult for a large company to develop for a Linux crowd in general.

Re:All Software is complex.

[IAmTheDave]

Or put another way, a guy from Microsoft, who has probably never configured or operated any of the systems he mentions, is telling a group of people, who also have probably never used those systems, that it's really scarey if you move away from Microsoft...

I don't think it's this bad - I think he has a point (that he puts his own MS positive light on) - but OSS is very often written to solve the problem a developer has, and is then supported and primarily used by developers. Setting up the software isn't something that the tech-savvy are concerned with because I think there is a very large "it works for me" chip on the shoulder of the OSS community.

That has to be one of the only reasons a good graphical installer for Linux doesn't exist today. I'm even dissapointed in Ubuntu in that light - they're the closest in my mind to a full desktop solution.

Now, that doesn't have anything to say about proprietary software, as the GP poster pointed out - often proprietary software is just as difficult to install, maintain, etc. I think OSX has it pretty much right with software installation - drag this "file" into your Applications directory. Done.

Isn't that better than MSI or make/make install?

Re:All Software is complex.

[budgenator]

Oh Bull, Linux isn't hard because it's Linux or sort of LSB compliant or anything else other than it's hard because its secure and because it's secure it's difficult to install libaries willy-nilly all over the file system; when Windows becomes secure, it'll be difficult too. Right now the windosers just don't get basic security from the inside, they only get it as a tack on from the outside; and running windows is becoming a huge pain in the ass; so much of their tacked on security is completely brain-dead nothing is integrated. Explain to me this why can't I download an application to install, save it to my desktop, rt. click the installer file and run-as admin to install it in the privilaged programs directory?

I find it silly that different distros keep configuration files in different locations I find it silly that in windows you can't, just seems elegent to me to have some system-wide defaults a normal user can't change and some stored in his user area that he can.

For the most part when I install a OSS program that has been ported over from Linux, it runs fine as an unpriveleged user, I can normally install it in my user area and not pollute the system files with beta or alpha quality software and it runs; compare that to installing McAfee, you have to use IE period, you have to let McAfee install an ActiveX controll into IE period, you have to do this as admin, period and that means find site as user, find out you have to be admin, change user, wait fifteen minutes for all the autostart and ET-Phone home programs decide to let you use your computer! Then after you've got it installed and running guess what a window pops-up say "new definitions found, you need to be admin to install"! In linux I can just

su -c"pacman -S clamav"

type in the admin password and pacman goobles up everything I need and installs it. In windows if you try to install an windows OSS program or a commercial program anywhere but the default location you're asking for trouble.

Re:All Software is complex.

[init100]

I find it silly that different distros keep configuration files in different locations, use different init scripts, use different install methods, have varying level of compliance with the LSB, are focused on either Gnome or KDE predominately, etc.

Maybe because people are different?

Since everyone can take the available pieces of free software and assemble them in any way they like to form a complete operating system, would you really expect everyone to be essentially the same? People have differing preferences and different solutions to various issues in building a Linux distribution. The major distributions may work together to create standards för certain parts of the operating system, but don't expect everyone to come along.

As long as there is freedom, people will make different choices.

Re:All Software is complex.

[ozmanjusri]

a guy from Microsoft, who has probably never configured or operated any of the systems he mentions, is telling a group of people, who also have probably never used those systems, that it's really scarey if you move away from Microsoft...

And this is NEWS?

It's even sadder than that. If you have a look at the website of IBS Synergy, the ISV they're quoting, it's an amateurish effort, full of spelling errors and broken links. The company has a grand total of five customers, two of which seem to be the same organisation, and one of which appears to no longer exist.

If this is the most authoritative source Microsoft can assemble to substantiate their claims Open Source is complex, I'd say they're a long way from being convincing. It's almost sad to see they're still stooping to such pathetic tactics.

Re:All Software is complex.

[omyar_hunt]

The complexity cost of a system will be as high as the competence level of your development staff. Open Source is complex in that is favours good, solid programming over lazy inept programming (though the perpetual global review and analysis that goes on in any healthy OS project).

However, much of the complexity is taken away after a certain threshold of competence is acheived.

In my experience, when working on a problem in a windows system (development or support) there is always confusion around what has cause the problem - is it video drivers, is it this dll, is it something else entirely. Cost of complexity and the vast unknown that is the Windows system is huge.

On the flipside, a linux discussion around a given problem contains less confusion about what caused it, and more suggestions on how to fix it.

Re:All Software is complex.

[dhasenan]

Moreover, if someone releases a binary for RHEL, you'll find patches to make it work with any number of other distributions.

The communities support your software for you. Nice, isn't it?

Re:All Software is complex.

[init100]

That has to be one of the only reasons a good graphical installer for Linux doesn't exist today.

Then what is a good graphical installer? The Windows XP installer? I'd prefer the Fedora Core 5 installer over the Windows XP installer every time. It is not only the graphics, but certain other issues. One prominent example is that the Windows XP installer won't ask everything before the installation and then work for half an hour. It will work for five minutes, ask a few questions, work for five minutes, ask a few questions, ad nauseum. The FC5 installer asks everything needed for the installation before it starts working.

But of course, most users never install Windows, since it comes with their computer.

Re:All Software is complex.

[Enderandrew]

Your response makes little to no sense. I'm talking about developing. I didn't say that Linux is too hard for an end-user. Developers enjoy standards.

I never said Windows was more secure, or better on the whole. So don't place such things in my mouth.

I said that because of the complexity of the GNU/Linux world, it is more difficult to develop an application for a broad user-base, and that is the truth.

Re:All Software is complex.

[Enderandrew]

That's why the LSB exists. You can decide what packages you want, how you want to compile them, and how you want to configure them. You can customize your kernel, and even choose from a plethora of file systems. Linux supports massive servers, cutting edge technology, and small embedded devices. If every distro followed the LSB, you'd still have a great deal of freedom, but developers could write code and know that it works out of the box on every distro.

However, that isn't the case.

If you write for the Windows API, then you know it will work for every Windows user.

In the GNU/Linux world, we can't even agree on one standard for distributing packages and installing them. For better or worse, the LSB states that packages should be distributed in RPMs. And while RPMs are extremely common, many distros openly oppose them.

If I'm a developer, I don't want to have to release a Debian package, a SUSE package, a Gentoo package, etc.

Re:All Software is complex.

[foniksonik]

Just to be fair.... OS X apps really do the full install the first time you run the app... ie: they copy all the support files they need to the Library/Application Support/ dir and generate preference plist files as needed, etc. Registration of the product is handled after the copy as well... etc. So it's not just a drag/drop operation. Apple does however understand that if you break up configuration into multiple time-shifted events, then it seems like a much simpler operation and of course since most of the config happens behind the scenes without dialogue boxes, etc. the user doesn't even realize it's happening... which is good and bad.

OTOH... more complicated apps do in fact have installers... and apps which want to present a EULA or a quick start guide also present a dialogue window at the very least.

What's nice still is that simple applications that should be easy to install ARE and more complicated apps that you really don't want to believe are that simple to install DO have installers.

Re:All Software is complex.

[Anonymous Coward]

You should check out InstallJammer (http://www.installjammer.com/).

Re:All Software is complex.

[mrchaotica]

I'm all for freedom, but I find it silly that different distros keep configuration files in different locations, use different init scripts, use different install methods, have varying level of compliance with the LSB, are focused on either Gnome or KDE predominately, etc.

It does make it more difficult for a large company to develop for a Linux crowd in general.

Yet somehow, all the Free Software manages to cope with the differences just fine. Why is that?

Re:All Software is complex.

[swillden]

It does make it more difficult for a large company to develop for a Linux crowd in general.

Somewhat, but it's not that bad. I wrote a commercial, closed-source Linux app that had to run on multiple distros and it's really not that much of an issue. The app I built was one of the more difficult ones to support across distros, too, since it had to integrate with (or replace) the login process, screensaver, etc. These are areas where distros do things very differently. Normal applications have many fewer issues.

Some things I learned:

• At core, all major, modern Linux distros look the same. There are small differences in they way files are organized and larger differences in how packages are installed and managed, but these differences don't require much effort to work around, even when you're changing the most distro-specific elements of the system (boot processes, login processes, etc.). Liberal use of configuration files and glue scripts is a very good idea so that you can reconfigure for lots of different environments without changing the binaries.
• Cross-distro development is easy. Cross-distro packaging is fairly easy. The hard part is the cross-distro testing, not because it's particularly difficult, but because you have to do it on every distro you're going to support. It's well worth automating as much of the testing as possible. I recently began trying to use a test-driven development process, and I wish I had used it on that Linux project.
• Although it's possible to create a cross-distro installer, your customers will be much happier if you provide native installers that integrate properly with their package management system.
• You cannot expect to make a single RPM package for all RPM-using distros. The RPM SPEC files won't differ hugely from distro to distro, but they will differ if you want seamless integration.
• For a corporate app, you really only need to target a small set of Linux distros to cover nearly all of your market. Red Hat, SUSE and Debian cover it. In the case of Red Hat and SUSE you need to support both the current release and the previous release. For Debian/Ubuntu, it's less clear. Ask some customers and take a guess (and see my next point).
• You can treat the too-many-distros problem as an opportunity. In my case, I recommended that we do three things: First, create packages that cover 95% of the target market. Second, provide a tarball with all of the binaries and some instructions on how configuration files need to be set up so that very technical customers (common in the Linux market) can figure out how to integrate it into the distro of their choice -- with the caveat that you will only support them in resolving issues they can reproduce on a supported distro. Third, (here's the opportunity part), offer to make it work in whatever distro they like, for an integration fee. Charge them enough that you can do a thorough job, including writing all of the automated test code for that platform, all of the documentation needed by your customer service department, etc., so that you can just add it to the list of "supported" distros, and make a profit doing it.

Really, I think the biggest difficulty with selling commercial Linux apps is the relatively weak demand. Although I don't like Windows, it's still quite dominant, and Windows apps are almost guaranteed a larger market. If, however, you can find a niche where there is significant demand for a commercial Linux product, the multiple-distros issue isn't going to significantly increase your development cost, will perhaps double the cost of developing you installation packages, and (assuming you make good use of automated testing) will probably increase your testing costs by 10% or so. Net, I'd say it costs <5% more to develop a significant application for multiple Linux distros rather than just one distro.

Re:All Software is complex.

[cl0r0x70]

". . . making server OS's look familiar tempts people to fiddle." Fiddling is good, and it's how most good engineers I know (including myself) learned the bulk of their knowledge. A good OS should encourage fiddling, IMHO. Being so scary as to discourage fiddling is a "bad thing." People just want to know that their inevitable fiddling won't likely break stuff.

Re:All Software is complex.

[rjstanford]

I find it silly that in windows you can't, just seems elegent to me to have some system-wide defaults a normal user can't change and some stored in his user area that he can.

That's exactly how well-behaved Windows applications should work. You write system-wide information into c:\documents and settings\all users\application data\[company]\[application] and user-specific information into c:\documents and settings\[username]\application data\[company]\[application]. Of course there are environment variables that the system sets for you in case the admin has changed the location, so you don't blindly write onto c: when you shouldn't.

Re:All Software is complex.

[Anonymous Coward]

That has to be one of the only reasons a good graphical installer for Linux doesn't exist today. I'm even dissapointed in Ubuntu in that light - they're the closest in my mind to a full desktop solution.

Do you smoke crack or something? Did you not notice the ugly blue/white text only screens that you MUST got through to install Windows NT/2K/XP/2003?

No, I would say Linux is WAY beyond Windows in capability during installation. Not only do most linux distros fire up X to provide a graphical installer, but you have A LOT more controll over the installer! With Windows installer your only option for passing arguments along to the installer to change installation options is to use the damn function keys. Example, pressing F6 to load different SCSI drivers before the windows installer goes through the worst possible method of trying to auto detect controller cards: throw every freak'n driven into memory and just see what sticks. Problem with this is the probing some drivers do can lock up some SCSI controllers before the driver for that controller loads up, so when the correct driver finally gets loaded the card is in an invalid state and the driver cannot find it. I have had to boot off NT/2K install floppies MANY TIMES because of this! With the a Linux installer you can pass arguments along right away at the boot prompt. And Linux installers PROBE the system to see whats out there BEFORE loading the device drivers! Instead of the half assed method MS uses. And even after a Linux GUI based installer routine starts you can still access a basic shell in most cases to see whats going on in the background and hack things that are acting funky. This is light years ahead of Microsoft.

Of course you have to have a clue to install Linux, which is why all these MCSE dorks are whinning and complaining. They want a cookie cutter approach to everything that is IT. Sorry, but the world doesn't work that way! I think these people are just upset because they feel the end coming and know their MCSE cert isn't going to mean dick in the new IT world...

Re:All Software is complex.

[Sylver Dragon]

You've not used Windows much in the past 3 years or so, have you?

Explain to me this why can't I download an application to install, save it to my desktop, rt. click the installer file and run-as admin to install it in the privilaged programs directory?

Ok, you can't do exactly that. It's more like:
download an application to install,
save it to my desktop,
Shift rt. click the installer file
and run-as admin to install it in the privilaged programs directory

I know, having to hit shift to have the 'run as' show up in the context menu is such a pain in the ass.

...it runs fine as an unpriveleged user...

You do, of course, realize that this is the fault of the program's developer and not Windows, which creates this problem. If a program is properly written, it never gets outside the HKCU registry hive during runtime; and certainly never touches system files. But, since we're bashing MS, it's ok to ignore the source of a problem.

installing McAfee, you have to use IE period, you have to let McAfee install an ActiveX controll into IE period, you have to do this as admin...blah, blah, blah...I don't know Windows well enough so I'm just going to bash it. Someone call the Wahhhambulance.
In linux I can just:
su -c"pacman -S clamav"
type in the admin password and pacman goobles up everything I need and installs it.

And I assume:
runas /user:administrator "c:\Program Files\Internet Explorer\iexplore.exe http://mcaffe.com/installer"
Is too hard? Granted, it would be eaiser if the path to iexplore was in the PATH environmental variable, but that would require some work on the user's part.

Windows does have a lot of problems (BSOD on program crashes, easily circumvented security, etc), but let's at least be honest about it.

Re:All Software is complex.

[Anonymous Coward]

Hey man, stop putting down my Pont of Sale System supplier!

Re:All Software is complex.

[Enderandrew]

Actually many developers do complain on a regular basis. And even more don't bother to develop for Linux in the first place because of the issue.

Let's say there are X number of dollars that big companies are willing to spend on OSS development in a year. And there are Y man hours.

Are we better served by having 50 million distros that aren't LSB compliant, or are we better served in developers aren't wasting dollars and man-hours catering to each distro?

And in the end, they're only going to release say 4 packages anyway, so many distros end up putting in time from their end to customize the software for their distro anyway.

Re:All Software is complex.

[cp.tar]

Now, that doesn't have anything to say about proprietary software, as the GP poster pointed out - often proprietary software is just as difficult to install, maintain, etc. I think OSX has it pretty much right with software installation - drag this "file" into your Applications directory. Done.

Isn't that better than MSI or make/make install?

Well, I don't know... I find "emerge prog_name" very convenient.

And for the most part, I need never fiddle with install CDs.

Re:All Software is complex.

[waveclaw]

The F/OSS Way is just different than the Microsoft Way(tm): the reboot, reformat, reinstall dance. Most problems in both OSS and proprietary cases are misconfiguration and misuse of APIs or resources like memory, bandwidth and scheduling. It just takes a little more training to solve the real problem than to have a grunt punch the power button every 30 days.

All software is complex. It is perhaps the most complex technology developed by mankind. (Almost) everybody that has walked on a bridge knows how bridges work at some simple level. You can't say the same for a web browser or video game. Completely virtual, arbitrarily defined and so abstract it takes years of study just to understand the basics.

With all ISV's it's a select your platform game:

1. Pick your proprietary OS (Windows 2000, 2003, XP Home,XP Pro).
2. Pick your proprietary DB(MS SQL Server, SQL Server 2003, Access, Oracle).
3. Pick your proprietary UI toolkit (COM, ActiveX, .Net, Qt).
4. Pick your proprietary webserver (IIS, PWS (I'm not kidding on this one),IBM Weblogic).
5. etc...

Is this less complex than OSS? No. Just different.

OSS is very often written to solve the problem a developer has, and is then supported and primarily used by developers.

And proprietary software is written to make a quick buck. There is software you can put in your datacenter in which 70% of the code is dedicated to license enformcement: copyright checks, license servers, date bombs, vendor backdoors and product key checkers. I installed a zip library on a system in the mid 90s that was about 2k in size, but the software to enforce product key cheking was over 50k. They did wonderful Q&A on the installer to make sure nobody cheated on the product key, but screwed up the library. Had to get a patch for the library from the vendor back in the day when you had to get CDs and disketts of patches shipped.

That has to be one of the only reasons a good graphical installer for Linux doesn't exist today.

Please define 'good.' If you are using a graphical installer, you aren't in a large server envrionement[1]. In data centers and computer rooms, standardization is so important that jumpstarting Solaris and kickstarting RHEL is required. You define what is your platform for 100+ servers then enforce that through automted installation. No WIMP-iness here. These are text mode and command line tools: drop in a CD and let'er rip methods of standardizing your platform.

M$ gets away with using graphical installers on their desktop-OS-sitting-on-a-server becuase you get squat for software on the installation disks. There is nothing to verify when a complete bells-n-whistles install includes the base OS and Minesweeper. You have to get extra 'software X' to make Microsoft platforms usefull, so your configuration boils down to MS+software X when talking to a vendor. This is no less complex than the OSS model. For support puropses, it might be less complete with OSS since the software is all from (supported by) 1 company (the distro maker) and all on the same disk(s.)

Why do people focus on the installer anyway? With OSS you shouldn't be needing to reboot/reinstall to fix things. Working in a larger datacetners, you will focus on the disgnostics your can get out of a program to manage it. The quality of installers doesn't matter too much unless you are doing lots of turnover.

Now, that doesn't have anything to say about proprietary software, as the GP poster pointed out - often proprietary software is just as difficult to install, maintain, etc.

I aggree whole heartily that OSS solutions are at least as difficult as commercial. However, the point of the FUD in the article is that OSS is a babylon of platforms, not that the installers suck. From the article:

"You don't get that repeatable [de

Re:All Software is complex.

[mrchaotica]

I think I was unclear. The point I was trying to make is that, generally speaking, any given piece of Free Software is going to work on whatever distro you run it on, and the way it'll do that is by configuring itself (using Autoconf or otherwise) to the system-specific settings at build time. If proprietary software vendors simply distributed their products in source code form, there wouldn't be a problem.

Re:All Software is complex.

[gbjbaanb]

If this is the most authoritative source Microsoft can assemble to substantiate their claims Open Source is complex

No, this and this are the sites that say that. The OSS community doesn't need MS to point out flaws, when we can do it ourselves. The correct attitude to all this, of course, is to acknowledge valid points and fix them (because if you don't, well, that way lies Marketing)

Re:All Software is complex.

[dublin]

Because, lets face it - what Gavin is saying here is that proprietary software vendors find it hard to develop for linux. *shrugs*

No, I don't think that's what he's saying at all. But without question, development for Linux and other Open Source operating systems such as the BSDs *can be* much more difficult, simply because the environments themselves are not particularly well-defined, and there's very little in the way of services, libraries, daemons, etc. that can be reliably counted on to be present and configured in a predictable way.

Development per se may not be that much harder, but delivering a product that will work and be supportable in the customer's environment is very much harder.

This problem is not helped by the fact that people who try to address this problem are routinely savaged by the open source (especially Linux) community - Corel, Linspire, and lately, even the beleaguered Red Hat could be the poster children of this disdain. Linspire's Click-and Run repository is arguably the best *reliable* way to get and upgrade apps on Linux from the customer's point of view, but the Linux community refuses to embrace it, even though it represents a level of testing and quality that the community desperately needs. (

And no, apt-get/update/yast/whatever doesn't count - it's *way* too easy for an open source update tool to completely hose your system, leving it in a state that is extremely difficult for an expert to fix and impossible for the average user. (Hardly surprising, since it can't make any resaonable assumptions about the state and layout of the OS it's updating, either, so this is a bit of a chicken-and-egg problem.) Personally, when tools like this hose the OS, which happens all to frequently, it's almost always easier to reinstall than try to really fix it. True, I have to do that with Windows too on a roughly biennial basis, but at least I can get really good software there, and security updates can be had without me having to hunt them down daily and *understand* them before installing them. Unlike the "average user", I *can* understand all that stuff, but why the heck would should I *have* to?)

The BSDs are slightly superior to most Linux distros in theis regard, since prerelease testing is much more a part of the BSD DNA, and there aren't a zillion distros to start with. This is definitely an area where the diversity of the Linux ecosystem is a very, very bad thing.

Re:All Software is complex.

[kimvette]

Of course you have to have a clue to install Linux, which is why all these MCSE dorks are whinning and complaining.

Actually, it's more often EASIER to install Linux than Windows. Linux installers typically make intelligent choices for you when you accept the defaults, plus if you have to add a driver for Windows to see the HDD controller it get sometimes get a bit tricky, especially if the hardware vendor throws drivers for all variants on the floppy."oops, the first driver didn't work. Second one didn't either. What card is this again? WTF, there are four different revs of that model, and to find out which I have to pull the card out?"

And then you have the controllers which will work in the Server OS but will not boot the server OS, but will allow Home/Pro to boot (yes, I know, buy a real hardware RAID, but show me the mom&pop shop folks who are willing to pay for actual server hardware. . . ), further complicating the process.

Windows' installer is not simple compared to Linux's - or at least compared to the distributions non-elitists run.

Re:All Software is complex.

[kimvette]

What is the complexity of choosing, say, three to five "supported" distributions plus "best effort" for others versus Windows, which consists of five or six server editions and four desktop editions (soon to be at least seven desktop editions), along with iffy-quality device drivers?

Re:All Software is complex.

[rob_figlabs]

>Why do people focus on the installer anyway?

I guess for the obvious reason - that if the installer screws up, or if you screw up the installation because the installer is difficult to use, then you're nowhere. The software may as well not exist for you.

Re:All Software is complex.

[labratuk]

I think OSX has it pretty much right with software installation - drag this "file" into your Applications directory. Done.

Isn't that better than MSI or make/make install?

I'll pretend you said apt-get rather than make/make install because that makes it sound like you have some clue about what you're comparing.

Not even close. 'Just dragging an application' is not package management.

With Apple Computer Inc.'s system there is no update system. What happens when $release of software acquires an exploit? Expect your user to regularly check for updates for all installed applications?

There is no form of dependency resolution. "But that's what leads to dependency hell..." - No. Software has dependencies. That's a truth. Dependencies need managing. Ignoring dependencies == ignoring the advantages of this invention called the shared library. Multiple applications can use one shared library to save disk space & memory. However the most important attribute of shared libraries is that a bugfix or security fix for the library needs to be updated precisely once and all applications which make use of it get fixed. Install multiple 'bundled' copies of a lib and you have to rely on every software maintainer making a fix and then go around fixing it n times. Good luck getting a fix for that crappy piece of shareware you installed 8 months ago and the author has no interest in anymore.

Re:All Software is complex.

[symbolic]

I don't think it's this bad - I think he has a point (that he puts his own MS positive light on) - but OSS is very often written to solve the problem a developer has, and is then supported and primarily used by developers.

Probably so...but I don't see that as a bad thing. Consider it v.1, an implementation of something at a very early stage in its development. With OS, once an idea is off the ground, there are no constraints on which direction it can take - the ability to fork something so that development does focus on the needs of the user community at large, isn't something that can be offered by the proprietary software model. where you're pretty much stuck with what they give you.

Re:All Software is complex.

[Blakey Rat]

I'm guessing he's probably talking about software application installers, not OS installers. The only software installer I saw in Linux when I used it was the one for ATI drivers, and it was not only terrible but it didn't work on my computer.

Re:All Software is complex.

[jesterpilot]

I think OSX has it pretty much right with software installation - drag this "file" into your Applications directory. Done.

Modern Linux distros all have a "Just click the programs you like - we do the rest" system. That is easy and user friendly. All other systems suck. They're only good for EULA's and for people who want to sell software on CD.

Re:All Software is complex.

[Schraegstrichpunkt]

If you write for the Windows API, then you know it will work for every Windows user

That's bullshit. Have you ever written a piece of non-trivial software for Windows?

Re:All Software is complex.

[I'm Don Giovanni]

I think OSX has it pretty much right with software installation - drag this "file" into your Applications directory. Done.

Isn't that better than MSI or make/make install?

Apple disagrees with you. Apple's own software uses installation packages that launch installation apps just like MSIs do. When's the last time you installed an iWork or iLife app?

Re:All Software is complex.

[Anonymous Coward]

uSoft needs to maintain a continuous stream of FUD against OSS. When I think of having to deploy a friggin' dual-core Clustrino, just to run their graphical whore of an OS.... I cringle!!!! What good is Moore's law when you are never going to realize what that Xtra CPU power can do????

Now, take that same rig and run whatever dist. of Linux you want on it... you can be serving up blog pages to half the world while you're Gnoming on something else! Further, OpenOffice, with its sensible XML open format will allow more lightweight collaborative apps than .Net or any of their other pseudo 'Live' tools... tell me I'm a dork or what??

Re:All Software is complex.

[Jarth]

With complexity many things can be meant.

- Unknown code and unforseen approaches. Lack of documentation and uncommented code. Blueprint
- Indeed, lack of experience outside of the "click" behaviour.
- Complexity is also a matter of appliction-to-user interaction. From a non-developper point of view, open-source is somewhat "stuck in the lab", meaning the overal look of such an application is not eye-candy, to many too technical terms or technicalities are required. There's no easy way to solve this. Though this is recognised by FOSS given their efforts on this matter.
- Open-source thrives on expertise, most small business companies lack such 'expertise' by cause of "piont-n-click" mentality.
- Open-source is too many times viewed as a few applications and a kernel, it's a license after all isn't it ?

There's also a matter of perspective: Indication such complexity might be at hand are actually numerous. How many multi-platform applications, drivers, plugins are being released simultaneously ? Think of video-card drivers, flash-player, and probably many others. Though this does not mean complexity is at hand here the perception of this is created.

Running into unsupported hardware is another issue, though less of an issue the last few years. Still this impressions persists and is not incorrect by far. Wich is why i'd like to conclude stating an effort to get Microsoft to open up their driver library would be both insane but at the same time insanely cool, okay, just insane, still.

Re:All Software is complex.

[Anonymous Coward]

Proprietary software developers find it hard to develop for Linux for a reason, actually. And that reason is that proprietary software companies like Microsoft make compatibility between various versions of Windows (relatively) easy. No, this is not the same as the LSB.

Look, if you write software that runs on Windows 95 OSR1, there's a significant probability that it will run-- no modifications, no re-compiles, no changes-- on the latest Vista beta. Unless you were writing device drivers or system-specific utilities like anti-virus scanners, of course. But a lot of application software, including the stuff that a lot of smaller development companies work on, will work just fine, because Microsoft works for compatibility between versions of the operating sysem.

I'm not saying they're perfect at it. And I'm not saying that your favorite app from the Windows 3.11 era will run on Windows Vista 64-bit Ultimate edition. But as somebody who has written software on Linux and Windows, I can tell you that there is a significant difference.

Under Windows, I ship a binary, and it typically works on most systems. Sometimes, I have to distinguish between Windows 95/98/Me and Windows NT/2000/XP/2003.

Under Linux, I typically have to have multiple binaries for multiple distributions (not to mention the myriad packaging systems-- apt, rpm, tarball, et cetera), and there's always a statically-linked binary thrown in for good measure, because the same version of libraries like glibc (which should be a pretty consistent thing) are compiled with different (and incompatible) options. Oh, and the wonderful list of file system standards! /usr or /usr/local? How 'bout support for both? Then you have kernel version issues-- do you know that some major distros still ship with 2.4 as a default? Look at Slackware!

Frankly, developing for the lowest common denominator in Windows has a LOT fewer variables-- and a lot more tools-- than doing the same thing for Linux. Sorry to say it, but it's true.

I'm not downing on Linux or Open Source in general, here. The same things that frustrate me in this scenario are the same things that make me like it as a development platform: choice. I use OpenBSD and Linux a LOT. I use them because I am free to experiment, to customize, to work with what works best (this is one of the reasons that Linux is great for embedded development: you have that incredible ability to make things work exactly the way you need, but there's not as much pressure for portability to mainstream distros).

But claiming that claiming that developing for multiple versions of Windows is as difficult as developing for multiple distros and versions of Linux-- that just doesn't hold water.

Re:All Software is complex.

[init100]

I'm guessing he's probably talking about software application installers, not OS installers.

Ahh, I see. Then I guess one could compare the Windows application installers with the Linux package managers.

Re:All Software is complex.

[budgenator]

ok step one download install_flash_player.exe and save to desktop,
step two

Ok, you can't do exactly that. It's more like:
download an application to install,
save it to my desktop,
Shift rt. click the installer file
and run-as admin to install it in the privilaged programs directory

, windows XP Home SP2 (Dell OEM) responds:
with the Run As window and I uncheck Current user and check The following user: admin:password ->OK; which pops-up an error window;

Windows cannot access the specified device, path, or file. You may not have the appropriate permissions to acess the item

WTF admin doesn't have appropriate permissions? Anybody who sees the above as normal, expected, and uncomplicated computer behaviour is a lunatic; if the underlying operating system is this bizzar and convoluted, what chance does a program developer have to create simple elegent software? OBTH I didn't have to Shift Rt. Click to get the Run As option in fact the shift seems to be ignored.
If a program is properly written, it never gets outside the HKCU registry hive during runtime; and certainly never touches system files. seems that it's pretty rare that a program is written properly and installers have a habit of installing drivers in system areas, and putting crap in start-up areas and making everything preload and ET-phone home. it takes fifteen minutes for the average user to go from login to a useable desktop.
The article said OSS is too complicated, and I'm saying bullshit, there really isn't much difference anymore in the degree of complication and it's getting to the point where windows is more complicated than Linux. I think when Vista hits the streets people are going to look at the BSOD of Win9X as the good 'ol days.

Re:All Software is complex.

[budgenator]

Microsoft sacrificed security for user-friendliness which was actually quite astute back in the pre-network days for a computer intended to sit in the one user, one computer paradigm; Unix/Linux sacrificed user-friendliness for security which makes more sense in at multi-user networked paradigm.
Both systems are putting competitive pressure on the other and as a result the differences is rapidly decreasing. If Linux is too complicated for developers, whatcha gonna do when windows is the same? I don't think Linux is too complicated because an OSS program developed to run natively in Linux and port to Windows runs better and is easier to install than an OSS program developed to run natively in windows. The complication of developing and running windows is only going to increase because the old guard is getting the boot and a new generation is taking over at Microsoft and they are letting and or forcing Microsofties, the rank and file developers, to code up to professional standards, which means sooner or later a lot of ugly hacks left in for backward compatability are going to disapear once the speghetti starts to get straightened out.
Don't think the level of security effects the level of complication in program developement; what is going to happen when people's antivirus programs start to monitor a programs behaviour and your program is using depreciated functions and sloppy security allowed two generation of windows ago?

Re:All Software is complex.

[rufus t firefly]

If we are talking about software installers, people in the Windows world can keep their installshield wizards.

As far as I'm concerned, a good package manager is a thousand times more elegant than a bunch of installers which all get to have their own weasely way with your system, and never really uninstall properly.

And if it comes down to pretty graphics, there are tons of usable frontends for things like apt (kynaptic), urpmi (gurpmi) and up2date (I think it's called up2date-gnome or something similarly strange) which allow complete management and upgrade of system packages in batches without ever having to see a scary command line. I think anti-opensource people are barking up the wrong FUD on this one.

As for software being complex, Linux distributions can be either complex or simple, depending on which distribution and how you decide to use it. I know people who use Ubuntu and have never seen a command prompt, and others who use Slackware or Linux from Scratch and don't bother with a GUI at all ... it's all about choice. And not like the Ford Model T color options, either ...

Re:All Software is complex.

[Anonymous Coward]

"And no, apt-get/update/yast/whatever doesn't count - it's *way* too easy for an open source update tool to completely hose your system, leving it in a state that is extremely difficult for an expert to fix and impossible for the average user"

You do not, I repeat, DO NOT have the slightest idea what are you talking about.

"The BSDs are slightly superior to most Linux distros in theis regard, since prerelease testing is much more a part of the BSD DNA, and there aren't a zillion distros to start with. This is definitely an area where the diversity of the Linux ecosystem is a very, very bad thing."

And you are a troll too. And unkowledgeable bashing troll, you sir.

Mono

[StonePiano]

Delivering mono based software is straightforward. If the machines have a configured connection to repositories (like with Ubuntu's synaptic package manager), then install mono, and the new software just runs. In short, the virtual machine architecture simplifies portablity.

Re:Mono

[Whiney Mac Fanboy]

I think Java would have been a better example - as the VM is more mature & ported to more platform (last time I checked anyway).

Re:Mono

[StonePiano]

Yes, you are right about Java. It's more mature, and a better example of a portable VM... but it is not open-source.

...Although you can target open source platforms - that is you can install Java software on a Linux server or PC.

Re:Mono

[Anonymous Coward]

The source code is open. I look at it quite often to work out what the hell
they have done. Maybe you should have said it not using an OSI license. Which
you should know will change soon.

Re:Mono

[evil_Tak]

That has more to do with package repositories than with virtual machine architecture.

Take a C program. If the machine has a configured connection to a repository, then install the libraries on which it depends, and the new software just runs. In short, a good package repository simplifies portability.

Re:Mono

[StonePiano]

You are right that a good repository simplifies portability

But if I need to distribute my proprietory software to run on servers configured with various brands of linux, then I have a problem. My software is not part of the repository. I don't have teams testing its installation in a variety of environments.

However, the mono framework is tested like that.

So, while I can't count on much uniformity on the target system, I can count on the uniformity of the mono framework. Hence the virtual machine architecture. Now I don't care whether they are running redhat, suse or ubuntu...

Also, future changes to distributions could break my systems. Mono is better resourced to keep up with all that. So it is easier for me to just target the mono framework... and hope they maintain some backward compatibility.

That goes for Windows too

[cerberusss]

'Customers who run Linux could be operating in Red Hat, [Novell's] Suse, or even customized Debian environments,' he explained.

Customers who run Windows could be operating in 2000, XP, 2003 Small Business, ...

Solution: standardize. Where I work, IT supports either XP/Novell or Debian. If you divert, you're on your own.

Re:That goes for Windows too

[Simon80]

That's not standardization, it's exclusion.

Re:That goes for Windows too

[cerberusss]

True. However, I have the feeling that standardization sounds better :)

Re:That goes for Windows too

[Anonymous Coward]

I still use w98 you insensitive clod.

Re:That goes for Windows too

[Simon80]

But you just admitted that it's wrong!

Re:That goes for Windows too

[cerberusss]

I meant that it's probably more accurate. That doesn't make it wrong :)

Re:That goes for Windows too

[Simon80]

Eh, well, I think it's a misuse of the word, for the purpose of euphemising the fact that you refuse to support more than a few different platforms. I just don't like euphemisms I guess, hehe..

Re:That goes for Windows too

[cerberusss]

Actually, I'm a developer so I just live by the rules that IT sets. However, I think they're doing the right thing by excluding all but two platforms. You think differently about this?

Re:That goes for Windows too

[duerra]

I know I wasn't a part of the original conversation, but I agree and disagree. I understand the need, but I also think the problem needs to be fixed. Of course, the fix really needs to, IMO, come from a higher level - in other words, getting the Linux distros to agree on a standard, OR have an installation library developed that's made to target the different distros.

Re:That goes for Windows too

[Anonymous Coward]

Yeah, we'll I'm still waiting for AGP, 80GB HDD and USB support in Windows 3.1. Talk about insensitive clods!

Re:That goes for Windows too

[cerberusss]

Well, the different distro's each have a different focus. I don't really mind. On the other hand, for an independent software vendor this can be hell. Can this be fixed with a library? I doubt it... And as for agreeing on a standard: we have the LSB however, there's lots of stuff that's not covered (packaging). It's a problem.

Re:That goes for Windows too

[Simon80]

I agree, the vast amount of distributions are entitled to co-exist, and focus on whatever they want, but that also doesn't preclude some of them from focusing on compatibility for ISVs, pretty much what LSB is supposed to help with. And if I recall from my vague knowledge of the LSB, it does cover packaging, by saying that LSB software is packaged as an RPM, or a subset of RPM, I forget. So something like Debian can comply with that by supporting alien.

Absolutely. Unlike Windows where

[Flying pig]

You will only be supporting Server 2000. Or Server 2003. Or some custom locked down corporate environment. Or W2000 desktop. Or XP. Or XP SP1 (still having problems migrating to SP2.) Or SP2. With different flavors of IIS. And SQL Server 2000. Or SQL Server 2005. Or we have to use MSDE for this application because we only have SQL Server 2005 available and it won't connect to it.

Quite right. Microsoft has a huge advantage in terms of consistency and lack of complexity, provided of course that you just want to run Office on the desktop. Oh yes, which version of Office?

Re:Absolutely. Unlike Windows where

[db32]

You forgot to mention the Novell integration stuff...the ActiveX stuff...then there is the whole MS-Java VM vs Sun-Java VM... I assume you have never been to a trade show either...where every vendor is willing to sell you a different "Document Management" system for upwards of $10,000 that really is just a stupid crutch replacement instead of having admins actually MANAGE the file storage and keep users from saving crap all over the network where they don't need to be. I mean...these vendors can't even explain what the hell their products do half the time...I just wander over, ask a few basic technical questions and the market bimbos (I will never understand selling software with sex appeal) are filling my bag with promo junk to make me stop asking questions in front of other potential dupes...er customers. Yeah...the closed source world makes SO much more sense and is SO less complex. But hey...as long as its going strong I can go to trade shows and get bags full of free goodies. The best stuff always came from the vendor that could actually answer my tech questions...they were generally happy to have someone that could speak intelligently with them about their product and thus broke out the expensive promo stuff. :)

Re:Absolutely. Unlike Windows where

[Anonymous Conrad]

And that's why MS have humungous testing labs, and why they don't release 0-day fixes to security problems: they've got so many scenarios to test.

So in the Windows world you tell the customers which configurations you tested on and support. That's what the guy here is saying: the OSS world is too open-ended, it needs to slim down the number of platforms to cut down vendor testing and support burden.

Re:Absolutely. Unlike Windows where

[Diomidis Spinellis]

A while ago I compared the number of dependencies to other components between Mozilla and the Internet Explorer. I thought that the free availability of many open source components would result in a much large number of dependencies (and therefore complexity) in Mozilla than in the IE. It turned out that the opposite was true. One explanation could be that, because Microsoft isn't obliged to publish the interfaces of internal Windows components and maintain backward compatibility, Microsoft developers have an easier time in creating internally reusable Windows components. Of course, in the long term, this strategy will backfire, as demonstrated by the travails of the Windows Vista release.

Re:Absolutely. Unlike Windows where

[Anonymous Conrad]

A while ago I compared the number of dependencies to other components between Mozilla and the Internet Explorer.

Uh, I don't understand your process here. Most of the DLLs you highlight for IE are well-documented parts of the OS interface - all of the DLLs in the first diagram, for instance. To pick another example you include rsabase, crypt and msasn1 as IE dependencies - these are Windows's built-in SSL implementation. Yet you don't show an SSL library for Mozilla?

And, to be frank, who cares how many DLLs either program is broken into? Your focus was less-complex-is-more-secure but in both cases it only matters whether it renders the HTML you feed it securely.

Re:Absolutely. Unlike Windows where

[Anonymous Coward]

"Microsoft has a huge advantage in terms of consistency and lack of complexity, provided of course that you just want to run Office on the desktop. Oh yes, which version of Office?"

Unless you consider that a default installation of most any Linux distro (next, next, next, root password, user password, next) sets up the OS and OpenOffice.org (which is about 97% compatible with M$ Office, and that's about 95% more compatible that different versions of M$ with each other, IMHO) whereas you need to buy windows, install it (only one config to use...) and then buy and install Office....

Re:Absolutely. Unlike Windows where

[thePowerOfGrayskull]

This comparison doesn't even make sense. You've compared source dependencies to binary dependencies? That's silly. If you looked at binary dependencies for Firefox on Windows, you would find that it has 16 internal DLLs. And that's before any of the built in OS-level dependencies, many of which are the same as IE.

Re:Absolutely. Unlike Windows where

[Diomidis Spinellis]

True, the DLLs in the first diagram are documented, but some in the last one aren't. Search for example for documentation on msoert2.dll or msasn1.dll on Microsoft's MSDN. The external DLLs and open source package dependencies do affect security, because an implementation error in them can create a security problem in the browser. You're right about not showing an SSL library for Mozilla, it wasn't listed as a dependency in its port, because it is part of the operating system base platform. A similar case could be made for some of the IE components.

Re:Absolutely. Unlike Windows where

[Diomidis Spinellis]

With shared libraries most package-level source code dependencies end-up as binary dependencies in the distribution, so the comparison is not as silly as you make it out to be. The idea of comparing Firefox on Windows with IE is neat; I'll run a test, after I press the submit button.

Re:Absolutely. Unlike Windows where

[mrchaotica]

every vendor is willing to sell you a different "Document Management" system for upwards of $10,000 that really is just a stupid crutch replacement instead of having admins actually MANAGE the file storage and keep users from saving crap all over the network where they don't need to be

No kidding -- the company I work for is implementing SharePoint for that very reason, completely screwing over those of use who use anything other than Explorer to interact with the shared directories.

Re:Absolutely. Unlike Windows where

[larpon]

Oh yes, which version of Office?
Office Vista?

Re:Absolutely. Unlike Windows where

[Diomidis Spinellis]

I just finished comparing the sizes of the DLLs the two programs use. The total size of the DLLs that the Internet Explorer loads is 44MB; the size of the DLLs for SeaMonkey is 11MB less: 33MB.

Re:Absolutely. Unlike Windows where

[Sylver Dragon]

Having inherited a Sharepoint implementation, I've found that the best way to deal with the browser issue (assuming you use FireFox) is the 'IE in a Tab' extension. Sure, it's still IE running, but at least it integrates into your FireFox browser, and with some rules work on the extension, you can make sure that your Sharepoint site links come up correctly.
And yes, Sharepoint et al. are really just a way of getting everyone to organize better. Though, the document workspaces are nice, since they offer automatic versioning. Also, the real pain with Sharepoint is that it pretends like it's going to integrate with AD and Exchange, but that's largely a sham. It knows they exist, and can even get usernames from them, but that's about it. Because it was a business decision to move to Sharepoint, I'm stuck with it, I'm just hoping that the next version is at least half as nice as MS is promising.

Re:Absolutely. Unlike Windows where

[mrchaotica]

'IE in a Tab' extension

I discovered that just a few hours ago, actually! (Or rather, I knew about it before but didn't pay attention because I don't use Windows at home.)

And yes, Sharepoint et al. are really just a way of getting everyone to organize better.

The trouble is it screws up those of us who knew what we were doing to begin with. For example, I use a bash shell (via Cygwin) half the time to access the shared files, and I can't do that with SharePoint. Another developer has a backup shell script, which won't work with Sharepoint. Etc.

If you think about it, every problem SharePoint "solves" already had a solution before, and in most cases the solution was simply a filesystem. SharePoint is just a crutch for people who don't understand how to use the tools they already have (or, from another perspective, a workaround for Windows' crap UI).

Re:Absolutely. Unlike Windows where

[db32]

The only thing SharePoint and other such nonsense packages "solve" is the middle management justification for existance by making themselves look busy and important in their reports to senior management about their integration of interoperable enterprise system management document revision organizational structure buzzword buzzword buzzword buzzword.

With the original point...how the hell canthey possibly say open source is less organized? I mean seriously, every closed source enterprise out there that has a non technical guy holding the checkbook is a hodgepodge of these strange 3rd party nonsense packages.

Re:Absolutely. Unlike Windows where

[thePowerOfGrayskull]

I appreciate the time you spent looking into it; the only flaw I see is that the acroreader dependency should be disocunted -- b eceause if you don't have Acrobat installed, it won't exist -- and it's not an MS product. If we assume that the size is larger because there's more compiled code (most likely a safe assumption), then by these results it's definitely true that IE is more complex in terms of dependencies.

On the other hand, I'm not convinced that this is bad thing. (BTW, avid firefox & linux user & developer here... just for the record.) Those same DLLs used by IE are also used by many, many Windows applications. There -- theoretically -- are no concerns around versioning, and a developer [assuming he develops only for XP SP2+] is virtually guaranteed that all of the dependencies he's using are going to be installed on client workstations.

When developing for Windows vs Linux (before I gave up and just switched to Java), a huge headache was managing dependencies and versions for Linux. Sure, the problems existed in Windows -- but to a much lesser extent. As far as I've seen, these kinds of issues are the single biggest reason why there are no consistent standards defining how to install and manage software across Linux distros. (And don't get me started about LSB.)

Open Source Finesses

[TaoPhoenix]

At first glance it seems to me that Open Source is indeed more difficult, because it is guided by different principles than Microsoft. Also, with less centralized money, individual variants have their own quirks.

Learning curve

[Libor Vanek]

Generally UNIX systems have much steeper learning curve - it takes longer to learn how to use them but then the user/admin/developers "performance" is much higher. So on the 1st or 2nd look it can look complex but once you start thinking how would YOU design these systems to "fit", then most of the time you'll come to the same solution as already exists - think PAM, X11 (protocol is crap but design is good), pipes, all-config-in-text-files etc.

Re:Learning curve

[Anonymous Coward]

All config-in-text-files is the biggest piece of junk, since every software uses its own format,stores it in a different directory and you need a different parser for each config file. A global registry like in windows is the right way to go (allthough the windows registry editor is crap).

Even gnome realized this and stores their config data now in a registry like directory structure.

Re:Learning curve

[SkunkPussy]

"X11 (protocol is crap but design is good)"

I'm very interested in this (throway) comment - how do you mean that the protocol is crap but the design is good? Surely if the protocol is crap, the design of the protocol is crap? Or do you mean that overall the seperation of client/server is good, but the network protocol is too chatty and thus crap? or what?

Re:Learning curve

[Libor Vanek]

I'm no expert in X11 protocol but AFAIK the protocol needs "confirmation" for every single "command" - which makes it extremely sensitive for network latency. Otherwise is great that you can run applications from different platforms and display them on your computer screen no matter what you got - Windows, Linux, Solaris... Also the idea of extensions which makes it not-so-difficult to add new things like transparency, shadows etc. is quite good (again - I'm no expert in this)

Re:Learning curve

[backwardMechanic]

all-config-in-text-files

Hooray for that! Whoever thought XML config fiels were a good idea? 12 layers of nested tags before you even get to a parameter - not so easy to skim over. Plain text config file with some provision for commenting is the way forward.

Re:Learning curve

[Anarke_Incarnate]

Put down the kool-aid. The registry is a total crapfest. I would not mind a central directory with independent plain text readable config files with independent names (similar to /usr/lib) for each system (as an option).
However, one of the strengths of UNIX/Linux is that you don't have to have the user rights to install in the central areas. You can install into your home directory and not hurt other people when you make stupid mistakes (and users will make them, as will admins).

By the way, why would you need a different parser for each config file? It is usually plain text in different formats. Sometimes they are simple, sometimes they add.....XML, etc. It is not rocket science.

Re:Learning curve

[Anonymous Coward]

Or how about having one file per setting in a (normal!) config directory tree?

name and path = which setting
content = setting value

Oh how I wished...

Re:Learning curve

[Anonymous Coward]

I'm no expert in this

Indeed. :-) Actually, the chattiness is more a fault of xlib (libX11, the C binding to X11) than the X Protocol's design. So much so that there's a project making a modern alternative to xlib called XCB http://xcb.freedesktop.org/wiki/- still uses the exact same network protocol, it's just a better designed C API to it!

Re:Learning curve

[Anonymous Coward]

No, the X11 protocol does not need confirmation for every request; it's Xlib's design which makes it synchronous. The xcb library has been designed as an asynchronous replacement for it, and does not have that problem, while using the same protocol.

Re:Learning curve

[Blu-Ray]

what?

No, the windows registry is a menace, its an impossible big heap of mostly crap. any program puts config string with diff info everywhere. one Software Vendor can create any number of product tree's cause developers 'just' think of something else every time a product or update/refactor comes out and they don't seem to manage how its gonna be put in there.

information is stored in there never to be found again, and it grows, oh boy how it grows ;)

the windows registry is the worst invention ever, period.

Re:Learning curve

[cortana]

Until you need to deal with anything in an encoding other than ASCII. Or want to reference one file from another. Eventually the home-grown parser becomes more and more complex, evolving into an under-specified, buggy clone of Common Lisp. :)

Re:Learning curve

[Ash Vince]

I can see why some one like yourself who has problems with english would not like config-in-text-files.

(the correct wording would be "All config-in-text-files are the biggest peices of junk")

The nice thing about storing config info in a text file is that if it becomes corrupt a smart human (something you are definately not) can go through the text file and see where it is corrupt and fix it without needing any extra software. If it is stored in some sort of binary file you need some thing like regedit.

Re:Learning curve

[OhHellWithIt]

Generally UNIX systems have much steeper learning curve - it takes longer to learn how to use them....

I think it depends on how you define "use". My 10-year-old is getting along fine with Gnome under Debian. She doesn't have the foggiest about how it works, but she managed to find a bunch of games and is able to do her work using OpenOffice.

On the administration end, there have been a number of things that I've had to learn about, like how to get my daughter's machine to print to the color printer on my wife's Windows machine, and on my own Debian laptop, the wireless thing just isn't there yet for consumption by my wife. On the whole, though, Linux administration is far better than it was five years ago, and light years better than when I first did any twelve years ago. (OTOH, I guess this exhibits that there is indeed a substantial learning curve, if I'm still challenged with it after 12 years....)

Re:Learning curve

[admdrew]

By the way, why would you need a different parser for each config file?

You answered your own question:

It is usually plain text in different formats. Sometimes they are simple, sometimes they add.....XML, etc. It is not rocket science.

Who cares if it's 'easy' to write another simple parser to handle a different config file... the point is it's not universal, and therefore is a pain to manage.

I'm not arguing in favor of the windows registry, but I really don't config files in a hodge-podge of different formats helps anyone.

Re:Learning curve

[admdrew]

I can see why some one like yourself who has problems with english would not like config-in-text-files.
(the correct wording would be "All config-in-text-files are the biggest peices of junk")

If you're going to be a grammar bitch, you probably should've found something else to correct the poster on. It makes more sense that "all config-in-text-files" is a single idea; the plural 'files' portion of that isn't the subject, the whole concept is.

Re:Learning curve

[mrchaotica]

By the way, why would you need a different parser for each config file? It is usually plain text in different formats.

Err... you answered your own question. You need a different parser because they're plain text in different formats.

If everything were the same format (e.g. Mac OS X's Property List XML format) it'd be easier.

Re:Learning curve

[Blakey Rat]

Yes, but because it's more confusing, that means hiring people who know it is much more expensive for companies. Which is why Linux isn't taking over the server room...

Say your a medium-sized company looking to implement a new database server. You can choose a Linux distro or Windows, but you need to hire a sysadmin for it. You hunt the job sites and you find a Linux sysadmin for $90k, or you find 48 Windows sysadmins for $50k. Who do you hire? (Considering that the performance of both platforms is similar-enough that, honestly, it really doesn't make much of a difference.)

Re:Learning curve

[Anonymous Coward]

Hmm, if that is in response to my directory-tree-as-config, my answer is:
just use UTF-8 everywhere (sorry, but I really don't want the whole marshall/demarshall mess of object oriented programming, just with encodings instead) and symbolic links...

Re:Learning curve

[backwardMechanic]

Non ASCII is a problem, you're right. But I don't see why we need XML tags, rather than some simpler syntax. It could still be a standard, with a ref. implementation and all. Plenty of scripting/programming languages seem to have managed. #include seems to work well for including other files, and it wouldn't be too difficult to invent something a little more selective. Yeah, okay, I feel like Lisp is grinning at me. XML files seem great for fulfilling a general standard, but not to be all that human readable.

Re:Learning curve

[infolib]

The latency issues have been more or less resolved by FreeNX, maybe read this nice NX intro.

Check out cosmopod.com - I think they still offer free testing NX desktops.

Re:Learning curve

[Libor Vanek]

I'd definitively hire Linux sysadmin since I know that he'll be able to manage more servers and get higher uptime then Windows admin/platform ;-)

Re:Learning curve

[TheRaven64]

The Windows registry is a good idea, given a bad name by a flawed implementation. A better example is the NeXT user defaults system (now found in OS X). This has a standard interface and stores data in property list formats. Since the interface is standard, the storage mechanism is abstracted. OS X plists can be stored as NeXT's ASN.1-like format, XML, or a binary form. This is completely transparent to the developer. Even if they are stored as binaries, the user can easily translate each one into the XML form, and modify that.

Oh, this has been implemented by GNUstep too, so it can be used on any *NIX platform.

Re:Learning curve

[DamnStupidElf]

I'm not arguing in favor of the windows registry, but I really don't config files in a hodge-podge of different formats helps anyone.

I wonder why you'd want inetd.conf to look anything like ntp.conf or zoned.conf. All three configuration files do very different things, and the configuration files are tailored for each application. Configuration files that just need name-value pairs like ntp.conf or most other configuration files do just that with no added bloat. Programs that need tables, like inetd.conf, use them. Programs that need heirarchial structures, like DNS, use them. Some people have recommended using XML or some other formal format for all configuration files, but the problem is that there are still semantic rules for all configuration files that vastly outweigh the syntactic rules. It doesn't matter if you specify your zones in zoned.conf or some XML database, if you don't put the right data in them it just won't work. In my opinion, it's better to keep configuration files simple and very human readable, not to mention backward compatible, simply because the hard part is configuring it right, not just making it parse without errors. The more working examples a system administrator has to go on, the easier his or her job is.

Re:Learning curve

[kimvette]

Have you ever even looked at the registry?

No two companies store similar settings the same way. Oh sure, it's generally in a hierarchical format (pretty much by force), but look under each app's directory. Some vendors obfuscate settings, others maintain very organized and easy-to-follow keys, and others use a very flat structure with a mish-mash of understandable and cryptic keys and values.

Also, how do you send your config to someone else? Sure, you might be able to export the registry, but that's not workable with quite a few applications.

Re:Learning curve

[mvdw]

It makes me alternately laugh and cringe when I see my colleague use VNC to attach to a linux machine from his windows desktop. I respond in this way because it baffles me how a guy can have 3 1280x1024 monitors on his desk, then attach to a linux machine's desktop at 1024x768 and think there's nothing wrong. I have a dual-monitor setup on my desk, and use cygwin-X to connect to the linux machines I have access to. It's much nicer to be able to have windows worry about the window management. In fact, he prefers the VNC method, since it lets him use the linux menu (Fedora and ubuntu).

Eh?

[bloodredsun]

Sorry, but I read this as "Choice is confusing - stick with what you are comfortable with. Hey look, that's us!"

This sort of gibberish is what you would expect from the most popular product in the market who are being challenged for the first time in a while.

Re:Eh?

[MadJo]

hmm which Vista version do I want...
Yup, they really don't offer any choice, do they? And they aren't confusing their customers at all, right?

Re:Eh?

[Kennon]

Applying various shades of lipstick to the same pig does not a choice make. *cringes in preperation for the karma hit*

Open source is easier, if you know how to work it

[Eivind Eklund]

I am used to working with open source and I'm fairly used to working with proprietary software. I find it much, much harder to work with e.g. MS-SQL than either PostgreSQL or MySQL, as the resources around it is missing. MS-SQL has a ton of tutorials and screenshots and stuff for the proprietary stuff, but there's no working support community (I can't go on IRC and talk to a developer), the configuration systems are long and complex (probably because the developers do not use them to configure things and do not do support), and the use of graphical interfaces make the tutorials long and clumsy ("Now press "OK" in the dialog that had a screenshot just taking up half the page" is noticably more icky than "").

To add insult to injury, when the stuff is in trouble, I cannot go to the source code and find out what's up or fix that stupid error message that says "Cannot open file" but nothing about WHICH freakin' file.

In sum: I find open source much, much easier to support. When there's a problem, I can talk to other people that have had the same problem *and have had the resources to fix it*, unlike e.g. Microsoft support. (Microsoft support have actually called me to find out how to fix a problem in a Microsoft product - a problem that should have been trivial for them to debug if they had code access.)

Eivind.

Complexity and diversity

[Anonymous Coward]

Complexity is the downside of diversity, but diversity has important benefits.

I've got to agree... somewhat

[inflex]

Supporting -Linux- from a cold-start is a pain, not OpenSource.

With Solaris and FreeBSD (as examples) you know what you're in for when you get there. With linux you never quite know for sure. Sure, you can gear yourself up with most of the more common setups (Debian, RH, etc) but beyond that things fracture into thousands of variants. From starting scripts to configuration files, it's a mess.

Re:I've got to agree... somewhat

[elrous0]

The main problem with Linux and OSS that I've seen is poor documentation and poor support. This is not so important for the high-end geek community that really knows what they're doing--but it's a NIGHTMARE for less sophisticated users and IT managers used to Windows' simple "Plug and Play" approach.

And, yes, I'm aware that it's more like "Plug and Pray" with MS's security problems.

To the unsophisticated user; simply sticking in a disc, installing an OS or program is the de facto norm with Windows and proprietary software. They're not used to having to recompile their kernel just to install a piece of software, or being told there is no "manual" per se and NO phone number to call for support if they have a problem.

-Eric

Re:I've got to agree... somewhat

[nick.ian.k]

"To the unsophisticated user; simply sticking in a disc, installing an OS or program is the de facto norm with Windows and proprietary software. They're not used to having to recompile their kernel just to install a piece of software" Slow down there, Flash! What unsophisticated user is going to be doing any software installation requiring a kernel recompile? With the main distros out there and their pre-compiled kernels, such is only required to support more exotic peripherals and less common file systems. I think you'll be hard-pressed for such in most instances, and in an enterprise environment, where it *is* a bit more likely, the "unsophisticate" isn't going to be doing it, but rather the admins, who'd better damn well know what they're doing on a given system in the shop before touching it, 'cause that's their effing job.

Re:I've got to agree... somewhat

[labratuk]

They're not used to having to recompile their kernel just to install a piece of software

This is a fictional exaggeration that never happens. Stop perpetuating it.

or being told there is no "manual" per se

SuSE has always come with a very good dead tree manual. Ubuntu has a good guide and FAQ. The online manpages are the best manuals ever written.

NO phone number to call for support if they have a problem.

Novell, Canonical or Redhat will be happy to arrange such a support contract.

You are pulling this stuff our of your arse.

Re:I've got to agree... somewhat

[jimicus]

Have you used Windows lately? It hasn't shipped with anything on paper in years, and most suppliers don't support anything over and above "does the hardware work".

If you mean "The kid down the street doesn't know it", then yes, you may have a point there.

Zimbra

[jlebrech]

At work we have a project based on Lotus notes which has taken 12 months up to now

And its nowhere near to completion and cannot beat Zimbra or other. It's looks apauling and cost the earth.

And we bought this because Lotus domino was EASY to develop. Sure with the notes client on everyones PC, but this is web development.

So now which top 100 website uses domino for its backend heh?

How to deal with the 'complexity' of choice?

[Famatra]

Simple, have standards that people/distributions can *choose* to follow.
Projects such as Linux Standard
Base (and others, list / talk about them if you know of them) allows
distributions to have a common point and common environment.

I think GNU/Linux will learn from the fragmentation of Unix. Part of the
reason why Unix didn't develop a standard was that each closed version was
competing against eachother for customers. With Linux there is no to-the-death
competition since the work is shared and co-operative.

Re:How to deal with the 'complexity' of choice?

[jlebrech]

Yes, just look at xgl and aiglx. they are competitors but they have a lets-share code mentality.

Re:How to deal with the 'complexity' of choice?

[Anonymous Conrad]

Simple, have standards that people/distributions can *choose* to follow. Projects such as Linux Standard Base (and others, list / talk about them if you know of them) allows distributions to have a common point and common environment.

Sure but if you write your app for the LSB environment and 10 distros (say) claim LSB compatibility then you can't realistically assume it'll just work on all of them: you need to test it on all 10. Sure, that is the idea, but in the real world it's never quite that simple. (It's the Java write-once-run-everywhere myth: all the JVMs are written to the same standard, right, so why doesn't all Java code just work everywhere as intended? Because every JVM and environment has its own set of bugs and quirks. It's a great idea, it just falls down in practice.)

So you still need to do 10 times as much testing. And you still need 10 test environments set up so you've got the right distro at hand when your customer calls with problems. Etc.

Re:How to deal with the 'complexity' of choice?

[TobascoKid]

It's the Java write-once-run-everywhere myth: all the JVMs are written to the same standard, right, so why doesn't all Java code just work everywhere as intended? Because every JVM and environment has its own set of bugs and quirks. It's a great idea, it just falls down in practice.

Would you like to give some modern day examples, or is this just the "Java - write once, test everywhere"

Re:How to deal with the 'complexity' of choice?

[Anonymous Conrad]

Would you like to give some modern day examples, or is this just the "Java - write once, test everywhere"

That's what I meant. I wasn't trying to take a shot at Java, by the way: I just meant you'd can't rely on something to just work if you don't explicitly test it. But it is easy to find examples of things that still don't "just work" in all Javas, e.g. this eclipse bug from last month.

Re:How to deal with the 'complexity' of choice?

[admdrew]

IANAJP (I am not a Java programmer), but given my limited experience with Java and the JVM on different OSes, I understand it's not realistic to assume that all code will work *everywhere*. I really think it depends how much interaction your program is going to be doing with the OS; if you require a lot of system calls or end up interacting with the file system, it's safe to say that your code will not work 'correctly' on another OS. OTOH, if you're taking in and outputting OS-independant data, it's logical to assume that your code *could* be programmed once and run everywhere.

Re:How to deal with the 'complexity' of choice?

[a.d.trick]

The problem with LSB is that there's a number of things that it does poorly. Until it fixes these things, there will be a large number of distros that won't even consider joining. For example, the LSB advocates the rpm for package managment! Both the apt and the portage system have major advantages over rpm. If they can give us a sensible standard than maybe we will be more likely to follow it, but I'm not holding my breath.

stupid FUD

[grindcorefan]

Ryan Gavin's statement about Free/Open Source Software being to complex is a complete no-brainer and doesn't even remotely reflect reality.

The truth is, complexity in computers and computer software is a tricky thing to tackle. As computer systems become more complex, complexity itself becomes a problem as integrating all the different components of high-complexity software into a working system as a whole is getting more difficult.

The perfect example is ms's own windows vista. That piece of software is so complex ms just can't get it to work properly. Delays are the logical consequence. Otoh, Free Software profits from having the source code available, not necessarily reducing complexity but making it easier to get along with it.

Summary: Complexity is a problem for software, but it doesn't matter if that software is free or non-free. Ryan Gavins statement is just what you would expect it to be, a stupid piece of FUD that might sound somewhat sophisticated to a non-guru but every proper software engineer would be rofl about.

Software Engineering 101: software is complex

[Anonymous Coward]

Software is complex. Incredibly complex. Horrendously complex. All software. No exceptions. It's a fact of life.

Re:Software Engineering 101: software is complex

[lokiz]

"Software is complex. Incredibly complex. Horrendously complex. All software. No exceptions. It's a fact of life."

Right on target. I work at a primarily Windows shop and I can tell you from experience that Windows can be very very complex. The only real differences is that Windows appears to be prettier (aka- easier to use) at first. But once you get past the basics, your in the same position as an open source OS, if not worse. The open source OS just doesn't feel the need to pretend that they are simple. They give it to you straight, no bull.

Re:Software Engineering 101: software is complex

[anothy]

if this is what you learned from Software Engineering 101, you should go demand your money back for your entire education. you've learned all the wrong lessons.

yes, most software is complex, but it doesn't have to be. the complexity generally comes from a few areas, like legacy support and poorly thought out design compromises. compare, for example, the Plan 9 kernel, which is ~180k lines of code for about a half dozen architectures, to linux, which is... well, an order of magnitude more than that, at least, even stripping out the vast driver support. it's also better structured and more readable. then compare other components: plan 9's ndb with Unix's whole host of files in /etc (how many files contain some combination of hostname, ether addr, IP addr, and so on?). and that's just low-level stuff. move up the stack towards the user and it gets more and more true. Apple's Safari is such a great experience for most people who use it because it's much simpler than most of the alternatives, say IE 7. the land-line telephone world retains many of its customers because mobile phones are more complex to use. software doesn't have to be complex, and folks like you who assume it does produce most of the complex code, because you've given up. and once you give up on trying, sure, it all looks like it has to be complex. it's a nice self-reinforcing fatalist outlook.

sure, sometimes complexity is unavoidable. but we should strive to make that the exception rather than the rule. and it can be, if we put the effort into it.

Re:Software Engineering 101: software is complex

[NoOneInParticular]

true

Re:Software Engineering 101: software is complex

[renoX]

Plan9's simplicity is great (Limbo is nice too), too bad its GUI suck.
BeOS responsiveness (and order of magnitude better than Linux, Windows) was great also, too bad it's mostly dead.

Both have in common to have very little software unfortunately..

Re:Software Engineering 101: software is complex

[anothy]

you mean "its GUI sucks." or perhaps "its GUIs suck." you need to make the number match. and BeOS should be possessive. the final sentence fragment has a significant parse error. also, your comments are totally irrelevant, and presumes much.

Marketing twists...

[Woek]

Gavin noted that the flexibility of open-source software in meeting specific business needs also means systems integrators and ISVs have to grapple with complexity costs

Is he suggesting that a lack of flexibility in closed source (Windows?) is a feature?

Isn't it all too complex?

[thogard]

I'm finding most software I'm running today is far more complex than it needs to be. With open source you can look at the source code and maybe understand why its too complex but most of the time its just a developer taking a short cut.

We all have examples of complex software gone bad. I'm guessing the 1st open source example of this is sendmail 5. Its complexity was required for what it used to do and that ended up leaving lots of holes in lots of systems over the decades. For a while people learned from that mistake. IDA Sendmail cleaned up the config. Bind's config files were redesigned. CERNS web server was excessively complex and the developers of NCSA http learned lots of lessons from that. The Apache team learned from there mistakes. Today Apache 2 is much simpler in most cases that CERN's server was even though it does far more.

The major issue with complexity today is the confusion between an Operating System and an Operating Environment. Linux is an OS but Ubuntu is an OE. OS X has Mach as an OS but several OEs including FreeBSD.

I like the KISS (Keep it Simple Stupid) for daemons and opening systems. That means every step of starting the system should be clear and easy to understand. That means being able to read the config files (no binaries or unreadable XML please). It means that programs should use a limited set of shared libraries (Solaris init needs a buggy XML? why?) The OE can be as complex as needed but the OS should be simple and clean. If you forget that, your system is going to be owned by some script kiddie.

Of corse it is

[Anonymous Coward]

It can't be compared to Windows which is mostly a unified platform. This does not count service packs, which many big users reluctantly adopt because they can lead to incompatibilities (even if slightly).

Hopefully with the advent of LSB and Portland binary and desktop standardisation will be a non-issue.

Administration is a completely different set of problems, though. For now Windows administering easier to learn and documentation availability is better. In linux old style messing through myriads of config files and non consistent behavior accross distributions (example: LDAP+automounter, SSH authentication over Kerberos) still result in more expertise and more time needed to to set things up.

What do you mean Open source complexity?

[mrjb]

The complexity involved in writing open source is equal to that of writing closed source - the poster confuses open source vs. closed source with single-platform vs. multi-platform. Observation: A lot of closed-source software runs only on windows. A lot of open-source software *also* runs on windows. In this context, it is true that open source is more complex than closed source, because it is more flexible.

There aren't a lot of differences to expect in compiling a package under Redhat, or Suse, or Debian. The real trouble lays in hardware- and platform differences. I've written a cross-platform application that compiles and runs on Linux, Windows and MacOS. I'm quite confident that it will compile without changes on different Linux distros, but there was some trouble involved in getting it to work on Windows and MacOS. The added complexity is a small price to pay for having the flexibility of running your software *everywhere*.

To me, the benefit is that I've gotten used to common pitfalls, and I've picked up some best practices that *always* work rather than only under certain circumstances. It'd be a safe bet that on average open source programmers are more standards-aware than their closed-source colleagues.

There is no such thing as Linux

[TheRaven64]

This is one of the fundamental things that Linux advocates rarely get. From a user perspective, there are about as many differences between Red Hat Enterprise Linux and SuSE as there are between FreeBSD and OpenBSD, or Solaris and some other SysV variant. The fact that they have the same[1] kernels makes little difference.

You can run Debian with a few different kernels and, apart from differing levels of hardware support, there is no difference from a system administrators perspective. Code written for one POSIX-compliant will usually work on another if the same shared libraries are the same.

Saying you support 'Linux' is silly. It's like saying you support 'UNIX.' Saying you support RHEL makes sense. RHEL[2] is a complete operating system with a set of defined library versions, a documented filesystem layout, a minimum specified set of supported system calls, etc.

[1] Modulo a few hundred vendor-specific patches.
[2] Substitute your distro of choice here.

Re:There is no such thing as Linux

[Doctor_D]

If I had moderator points, I'd mod up the parent post.

Re:There is no such thing as Linux

[Anonymous Coward]

Well the kernel isn't even the same. If you run FreeBSD 6.x, anything written for that will probably run. In addition, you can run that binary on PC-BSD and DesktopBSD. With linux, there are so many changes between minor kernel releases that a 2.6.8 kernel might actually behave differently than another distro with 2.6.15. Plus there are vendor patches. Not everyone ships a vanilla kernel these days. For the most part its the same. An end user will rarely notice. An IT person who has a ton of boxes probably will.

You are absolutely right that people should state their distro. Compare redhat to gentoo. Gentoo does a few things very BSDish. The philosophy is totally different between the two. Like any small movement, they bring in other groups when it helps their cause and ignore them the rest of the time. Its very similar to the gay rights movement in the US. Sometimes gay men and woman want to stand alone and other times they include bisexual and transgender individuals as well. It depends if they need numbers that day.

Re:There is no such thing as Linux

[vhogemann]

Agreed,

Also, this mithic "Linux OS" hurts the image of all distros... Because it's too hard, because nobody supports it, because it's not ready for the desktop, because it doesn't support hardware X, because it doesn't play DivX movies out of the box, etc...

When some clumsy, half baked, buggy distro get's a bad review the press don't say "Clumsy Distro X sucks", they say "Linux sucks", and this is just plain wrong!

People must understand that Linux is a kernel, and must understand that RedHat is not Debian that is not Suse that is not Ubuntu that is not Mandriva that is not... ad nauseum.

Re:There is no such thing as Linux

[petepdx]

History repeats itself ? For those of us who where in the
trenches during the first Unix wars, know that history
repeats itself all too much. A friend at DEC told stories
of being told to ensure his tools would not work with
the 'enemy' camp.

Just as 50% of Americans are now convinced there were
WMDs in Iraq, what is the percentage of those that call
themselves technologically aware, that uSoft's is easy
to work with, uniform, totally upward compatible, yada,
yada, yada ...

Re:There is no such thing as Linux

[mrchaotica]

Saying you support 'Linux' is silly. It's like saying you support 'UNIX.'

Unless, of course, you actually do support "UNIX" in the way lots of Free Software does: by distributing the source with an Autoconf (or whatever) build script, so that it adapts itself to whatever environment the user compiles it in.

This should be the preferred way of distributing software, proprietary or otherwise.

If Open Source software is so complex to use.....

[QuietLagoon]

... why does Microsoft feel the need to spend so much time bashing it. Why doesn't Microsoft just let the simplicity, reliability, security and value of their own software speak for itself?

The problem that Microsoft faces, is that their own software does speak for itself, and is provides a convincing argument that Microsoft's proprietary closed-source model is failing.

Two answers

[rpg25]

1. Is open source more complex? Answer: No, consider SAP.

2. Is it hard to develop for Linux deployment? Absolutely. There's too much worry about what libraries are going to be available on a particular distribution. At the company I work for, we have pretty much given up development for distribution to arbitrary linux systems. Now we put our software onto a LiveCD using Knoppix, so that we know the platform will have the right utilities, libraries, etc.

Proprietary software *is* outsourcing:

[wild_berry]

Writing software is complex; maintaining an operating environment, too. Paying other people to do work you don't want to is outsourcing. I'm not yet been convinced that outsourcing does anything other than provide short-term gain and is often at long-term expense. And buying a software platform outsources the required development work. I'm not surprised that MS, as a software company, need to reiterate that software is complicated (because it makes people realise the value of their products), but for companies to expect that software development would be easy is daft.

Required Balmer Monkey joke

[Anonymous Coward]

I thought it was, "Developers developers developers developers"

This should be fun to watch

[Anonymous Coward]

This is the same argument MS used for Java for years. It isn't true for Java either but it amazing how many people ended up believing it even here on slashdot. The fact is most large systems are pretty complex. The art of writing software is giving the person who uses it the impression that it is very, very simple. Take Goggle for instance. The search engine does some truly staggering stuff to return results to you when you type them in that box. You don't generally think about this because it works.

Now think about how often when using an MS product you suddenly realise that it is doing some really clever stuff because it doesn't actually work. An example:
Excel used to have a limit on the number of characters in a cell. That limit was 255.
The limit was lifted, in, I think Excel 95.
However if you have a worksheet that contains text greater than 255 characters you can't copy the sheet to a new book. If you use the copy paste commands from the window bits of formatting get lost. If you use the "copy sheet" function Excel will truncate any bit of text longer than 255 characters. If you use the same function but perform a move instead of a copy it works just fine. Now that is complex.

It's about the same

[yancey]

Having managed both Windows and Linux systems in an environment with 500-1000 machines, I can say that the workload ends up being about the same. If someone were to tell me that managing Linux is "too complex", I would respond by saying that you just haven't yet learned Linux, but perhaps have learned a specific distribution. In essence, Windows is a single distribution and learning only one is easy enough. However, once you understand the fundamental concepts of Linux (or any unix-like OS), adapting to a new distribution is relatively easy. There is a learning curve with Linux, but there is with Windows too. Just ask anyone who has switched from a Mac to Windows. If you're not willing to learn, then you're just lazy.

Re:It's about the same

[Lumpy]

Windows is a single distribution and learning only one is easy enough.

really???

3 machines.... NT 4, W2K, WXP find the drive configuration dialog in control panel....

All 3 move it around like a cat and mouse game.

and YES you will find all 3 in a corperate environment still used today. there is no such thin as a single flavor you have to learn..

Opening a share under Server 2000 and server 2003 are different. server 2000 = wide open. server 2003 = closed down and you have to open every aspect.

the differences between 2000 and 2003 are dramatic and will confuse the hell out of a 15 year Windows IT vetran.

Re:It's about the same

[ericlondaits]

There is a learning curve with Linux, but there is with Windows too. Just ask anyone who has switched from a Mac to Windows. If you're not willing to learn, then you're just lazy.

The people that have trouble switching from Mac to Windows would probably never survive the switch from Windows to Linux, except for switching to a friendly desktop distro and never looking under the hood (hey! Just like Mac OS X! :) ). There are learning curves and then there are learning curves.

Lazyness aside, the fact that you should be willing learn doesn't make the switch easier. Sometime's it's not about advancing to a higher level as a human being, but about bussiness. I might be more than willing to learn about different Linux distros and become the "Master of FOSS Portability (tm)", but my boss might not be such as willing as employing my time thus (which he pays for).

Microsoft's argument is not directed to me, but rather to my boss.

Re:It's about the same

[miyako]

I agree with this in general. When working with software there are several layer of abstraction that one must understand in order to really be productive. Too many people spend time learning the specifics of one OS or one distribution, and are at a loss when it comes to using something marginally different. In essence, I think it's the same problem that we berate users for. A user learns how to copy and paste something in Application Foo, but then when they open up Application Bar they are at a loss, even though both applications use the same shortcut keys, and have the "cut" and "paste" options in similar places in similarly named menus. The same thing happens with developers and admins who learn how to set up a system by clicking the right options in one GUI, but lack a detailed understanding of the general principles that apply, and so have to relearn everything when they are put in a different environment.
I remember when i first started using Linux, it was back in 1997 or 1998- if you think Linux is rough around the edges now compare it to what was available back then. Before that I had only used Windows 95, and I had very little technical knowledge (I was in highschool, and I had just taken my first programming class. At the time, we did all of our development by telneting into a Debian server. I wanted to have the same OS at home, so I brought my computer in and my CS teacher helped me to install Debian on it). When I first started I didn't really know how to do anything. I didn't have access to an internet connection, so I couldn't search for help, and I was often frustrated because I didn't really understand at the time how Linux worked. Fast forward to two days ago, I just switched from Suse to Kubuntu. It's been years since I've used a debian based distribution, but the transition has been absolutely painless. Switching over wasn't a cause for any sort of stress, or really any sort of thought whatsoever (my old machine died, when I got my new machine put together I tried to install Suse 10.0 only to find out it didn't like installing on SATA drives, and I happened to have a Kubuntu disk lying around that I had been intending to try out at some point, so I just shrugged and installed it instead).

Re:It's about the same

[Dark_MadMax666]

Stop FUD please .If you cant figure out 2k3 after managing 2k you have no fucking right to be called "15 year windows IT veteran" .Heck you have no place whatsoever in IT .

Maybe training

[globalar]

Open source communities, including users, foster somewhat different attitudes and habits about software complexity. In a way, I expect an OSS package to show me its complexity up front. I want to know more about what my software is doing, how to configure it, etc. I don't "buy" OSS based on how shiny it is or whether or not the salesperson was attractive. And I try not to buy out of sheer ignorance either. So complexity - or someone's idea of it - is a definite plus. Much better than half-assed "simple".

For a Microsoft product, I fully expect that its going to balance stupid with too-simple. Thus its usefulness to me is significantly diminished. But I know people who are exactly the opposite, and look at a Microsoft product as keeping them from having to know things about their computer. Me, I'm more worried that I don't know enough ;)

Complexity is part of the beast

[jbarr]

At some level of computing, it's going to be complex.

I personally think that the "certification boom" really insulated many people from the reality that computing systems are, in general, complex. Many IT/IS people became so pigeon-holed that they came to believe that their little corner of the IT/IS world was all there was. Now, it seems that diversity of knowledge is again becoming the desired hiring trait over uber-specialization.

Besides, flexibility typically comes at the cost of complexity.

"complexity" == transparency(!)

[Anonymous Coward]

If they can't handle transparency ("complexity") they should go for alternative OS' like Windows or MacOS.

Linux is great for giving choices in every aspect covered by modern operating systems. You can configure Linux to do EXACTLY what you want, not more or less. If you don't care if the OS does EXACTLY what you want you migt stick to what suits your needs best. The market offers the right OS for everyone.

Short answer: depends on the user.

[Noryungi]

Long answer:

Is open source difficult? Yes, if you are just an average user. No, if you are a system administrator-type of user and that you manage information system for a living.

If you are just an end-user, someone who uses computer to do something else (creative work, accounting, marketing, sales, whatever) and you don't know anything about computer, then yes, I guess Open Source is still too difficult for you... unless you have a sysadmin close at hand to (a) install your machine and (b) make sure it's updated regularly. Then, Open Source can be -- should be -- just as easy (if not easier) than Microsoft products. Open Source GUIs, such as XFCE, KDE or Gnome, once installed and configured properly, are just as easy and friendly as Windows. Of course, the ultimate in user-friendliness is Mac OS X, but that's another story.

Please note that the term "user" -- as used above -- is not negative at all in my mind: I can perfectly understand that your job has nothing to do with computers, and that you don't have the time, or the inclination, to learn more about computers. And no, I don't think there is such as thing as a "Power User". Either you know enough to manage your own machines, or you don't. People who know just enough to be dangerous, but not enough to clean up the mess they have made, are users in my mind. Dangerous ones, but users nonetheless.

On the other hand, when it comes to system administrators, Open Source wins hands down. Things like Apache, vsftpd, NFS, CUPS, perl/python/shell scripting and, especially, OpenSSH make my life (and the lives of countless other people) so much easier than their Microsoft counterparts. Plus, they are a lot cheaper than all the Microsoft products, they are more reliable, easier to manage, upgrade, patch and install. Seriously, consider the following examples to upgrade a machine or an application:

1. Debian: sudo apt-get update && sudo apt-get upgrade
2. Slackware: sudo upgradepkg ./*.tgz
3. OpenBSD: sudo pkg_add -u -vvv -i
4. Etc...

Sure, to get to the stage that you actually can type these commands under OpenSSH and know what they do, you need to put in a lot of work. But the result is worth it. And, if you are a sysadmin worth his/her salary, you'll probably have a passion to learn that kind of things. Once learned, these commands result in less downtime, less cost, more customer satisfaction and a more efficient company. All in all, Windows, with its lack of security, Registry Database, its rather ugly GUI and its general flakiness is not good enough or "simple" enough when it comes to systems that must run 24/7 and support dozens, or even hundreds of users.

Linux, on the other hand, may not ready just yet for the desktop. But it will one day. Which is probably why there is an un-ending stream of FUD coming out of Redmond these days...

Re:Short answer: depends on the user.

[TobascoKid]

Much the same could be said for prorietary software (ie, Windows) - it's just that adminless users come accross the "complex" bits at different points in time. I have yet to meet a windows user who has not had thier computer become unusuable at least once - so they end up with the complexity of getting it back up and running. With Linux, you mostly hit complexity when you want to initially install Linux or new hardware, after that it pretty much just runs (and with Ubunutu, a lot of people aren't even running in to anything particularly complex, at least not for a while).

Re:Short answer: depends on the user.

[tehcyder]

Debian: sudo apt-get update && sudo apt-get upgrade
Slackware: sudo upgradepkg ./*.tgz

OpenBSD: sudo pkg_add -u -vvv -i
Etc...

Etc looks the simplest distro to me. Where can I download it?

Re:Short answer: depends on the user.

[Tony]

Seriously, consider the following examples to upgrade a machine or an application:

Ubuntu:

1. little icon appears in system launch panel thingy. Pop-up tooltip tells me I have 94 updates to install.
2. Click icon.
3. Enter my own password.
4. Profit!

Seriously, the whole upgrade thing has been a non-issue for a while now. Suse and Red Hat have had similar non-command-line tools for several years. It's not as easy to use as OSX, but it's damned close, and better than the MS-Windows method of upgrade.

I have installed Ubuntu on several computers for non-techies. It works very, very well as a desktop system. I bet the same could be said of Suse and Fedora, too, though I haven't tried either of them in a couple of years. (I'm a Debian snob.)

Re:Short answer: depends on the user.

[ProteusQ]

Linux, on the other hand, may not ready just yet for the desktop. But it will one day.

If it wasn't ready before, it became ready with the release of Ubuntu 6.06. I have been searching for five years for a desktop Linux that just works — as in Grandma friendly. I tried RedHat 7.3-8.1, Debian Sarge, Knoppix, and Kanotix. All of them had their issues unless you were at least a power user. Not so with Ubuntu, or my preference, Kubuntu. I'm not saying it's perfect or will replace standard Debian or Gentoo, but if K/Ubuntu isn't ready for the desktop, then neither is Windows.

i giggled

[QX-Mat]

I saw "opensource" and then "microsoft" and giggled.

I am, however, a tea-giggler. So now there's spillage every where :(

OS can be too complex, granted, but thats because where the light is focused, in development, there is a harsh learning curve. OS tends to be for OS programmers. Companies like RH, Mandriva et al, are the beautifiers.

Mind you, I wish autoconf never existed.

Wait a minute...

[Anonymous Coward]

..a notoriously anti-competitive monopoly is saying that options are bad?
That is funny on at least 17 different levels. LOL

That goes for Windows too-Stop picking on me.

[Anonymous Coward]

I know everyone's trying to draw an equivalency, but for the portion that people* actually use. Is there really that much of a difference?

*Keeping in mind the audiance that's being used in the story.

"Customers who run Windows could be operating in 2000, XP, 2003 Small Business, ..."

According to slashdot, the uptake of new windows is slow.* Very contrary to your "you could run into ANYTHING premise". In fact it's going to be mostly homogenious because businesses buy windows in bulk, be it from one hardware vendor, or the advantage program.

*So slow that some are running the 98' series, instead of the NT series.

Re:That goes for Windows too-Stop picking on me.

[cerberusss]

I know everyone's trying to draw an equivalency, but for the portion that people* actually use. Is there really that much of a difference?

No, but the article was about system integration companies. They hook into Windows products much more than your average user who browses, mails and IMs.

What a Crock!

[smartin]

Pretty much all implementations of Linux follow the same layout, /bin, /etc, ... Yes it's different from windows and yes there are some slight variations, but once you learn the concepts and discover that ascii configuration files are vastly superior to binary registries you will find that it is much easier to deal with. Most distributions also provide a set of simple and easy to use gui interfaces for configuration that make everything accessible to those who do not like to deal with ascii files. windows on the otherhand is a nasty complex beast that tries to hide any advanced configuration in layers of of ui and non-intutitive organization.

The bottom line is that it boils down to what you know, and whether you can be bothered to learn a different system.

Re:What a Crock!

[mojojojoe]

Yes, we are moving into the realm of desktop environments and I have something to say. Why isn't there a open source project to mimick the apple desktop. This could be based on Gnome. Why do I say this? I've tried to migrate two non techy friends onto Linux. The one was old, the other was artistic. They both came back screaming with frustration and asking to be put back on Windows ASAP. (I had used Ubuntu) Their difficulties: the one couldn't comprehend the unix filesystem standard. And why should he? I noticed that the apple file manager goes to great lengths to hide the unix filesystem underneath. It is actually quite a mission to get under the bonnet (for me a first timer on OSX). Secondly, the example of Windows should be followed by using lots and lots of wizards. You could then say, the more wizards a distribution has the more powerful it is. I also think that the choices are way to many. For the "dumb" users from Windows only the best software should be used, in a highly integrated manner. I think open source is made for this. Perhaps there should be a commercial venture to provide this though. The problem in achieving this is I think one of the open source rules "respect other peoples ideas" Okay. But if you are going to create a really fantastic highly integrated but easy to use system you are going to have to toss a lot out. The complete antithesis to all this is KDE. Sorry, I don't mean to flame. I have just been using KDE the last few days and I am really tired of it already. The problem KDE presents is the enormous amount of choice. I have spent days configuring everything, and at this stage of my life (I'm 34) I think it is a huge waste of time. There are nice aspects of KDE, but it is so untidy, I can't wait to log into Gnome again.

This just in..

[Rob T Firefly]

..Microsoft says exaggerated nasty things about their competition in order to sway people their way.

In other news, there will be weather today.

Re:This just in..

[ozbird]

In other news, there will be weather today.

I prefer the "complex" version with temperature, precipitation and wind speed estimates.

Re: "It's So Difficult"

[TaoPhoenix]

That sounds like a comment aimed at end users, because by definition server admins have the training to make their own choice.

To a point, for desktop End-Users, let's presume a simplified version of the Windows world where the company gave all the Win98 boxes to charity and all that's left is Win2000 and Win XP2.

With a very few exceptions, these two OS's are close cousins. "No, the media keyboard doesn't work as well with Win2000." But to a degree, once all the silliness goes away, there are relatively few OS level problems, and more user "I lost my file" problems.

From this viewpoint, having someone take away your Sloppy-But-Useable XP2 box and replace it with some brand of an OpenSource OS *WILL* cause pain. The terrified enduser will flee far, away, back to that comfy monopoly.

However, I am disturbed by the general reports of Vista. I think it will require significant work and adjustment for endusers. Suddenly, if the enduser *has* to struggle with something anyway, the choice changes.

"I'm going to be in for a storm anyway. Would I rather struggle with Vista, or take my best shot at a friendly development OS?" I'll freely admit I'll have a parallel XP box running as long as I can hold it together, while I slowly watch the market and see who solidifies as the Must-Have distro.

How about them apples?

[Anonymous Coward]

OS error.

/etc/foobard.conf

[Anonymous Coward]

I agree, controlling the foobar deamon by editing the plain text
file /etc/foobard.conf is much too complex. Oh for the simplicity
of the registry!

You would be amazed at the idiots they call IT....

[mikelieman]

Seriously. One of my favorites really thinks 192.169.1.0/24 is a valid private ip address.

Since the "Suits" perception of "IT Expert" is a Windows Monkey who is better than they are, needing to actually UNDERSTAND how it all works together isn't a requirement.

Re:Open source is easier, if you know how to work

[Anonymous Coward]

How about some examples? MSSQL is perhaps one of the easiest RDBMS to install/use of them all (yes, I've used Oracle, MSSQL, MySQL, and PostgreSQL) even if it doesn't scale as high. Put in the install CD and hit the various Next/OK buttons and you are done with a system that, by default, is pretty good. MS documentation is typically leaps/bounds ahead of what's available my man pages. MySQL actually has decent documentation for many things, but for others it falls on its face.

If administrators have to resort to the reading the source when something fails, there is a problem. This is at least as bad as having to put in a support call. Digging through the source of a project that you aren't familiar with (potentially written in a language that you aren't familiar with) is a big timesink. Figuring out and debugging why the failure happened in source that you aren't familiar can also be a huge effort. While it does give the specious feeling of accomplishment because you actually 'fixed it' yourself through your own effort when, in fact, you spent a whole week dorking around with something when you should have been able to ask someone or looked up the problem and have moved on to other projects after 15 minutes.

I'm not saying Microsoft is always the best but your examples frankly suck at trying to prove your point. You pick perhaps one of the easiest to install/use RDBMS and say it's hard and then you say that digging through foreign code in potentially strange computer languages to find some bug is the best thing since sliced bread.

The one point I *will* give you is that with OSS, you *can* do those things, but I'd only consider doing them in extreme situations or if I'm tasked by my boss to do so. With most software, OSS or not, if I'm not specifically tasked to get it working, if it doesn't work after 15 minutes, it's deleted and I find something else that'll do what I'm looking for. Life is too short to spend days trying to get someone else's piece of software going that coredumps or has some other wierd behaviour, especially given the amount of software out there these days.

Mod Parent Insightful

[kthejoker]

Especially because it's funny that you pay so many X dollars for proprietary support, and open source support is just so much easier to deal with. That's just sad on Microsoft's (and other software vendors - Oracle, I'm looking at you) part.

The guy on the street corner shouldn't have a better watch repair system than the guy in the jewelry store.

Re:Mod Parent Insightful

[Anonymous Coward]

Don't forget to include the salary of the admin in your cost for support. If a $100 call to a support center resolves a problem in 15 minutes vs. an admin person working on the problem for a week, which one is cheaper? That's time that the admin could be fixing other problems that are in the queue. Just because you will pay the salary of the admin anyway doesn't mean that the cost goes away. You have to go by the number of problems fixed per time and cost, not just "additional cost due to calling support lines".

What about static linking?

[vlad_petric]

True, doesn't work for everything (e.g. 3D acceleration), but otherwise does wonders with respect to compatibility.

Re:What about static linking?

[rpg25]

Thanks. Good point. Static linking does go a long way to solving the problem with respect to libraries. But other problems remain, notably the invocation of infrastructure services that are separate programs, rather than libraries. E.g., we have a program that uses ImageMagick to generate images for display through a web interface. Since the ImageMagick command-line has not been stable, making our program work across different linux distros, which have made different decisions about what ImageMagick versions to package, made us miserable. It has just been easier to give our customers full linux distros, which we know and control, rather than trying to fuss with each customer's individual set-up (or give them a directory tree with the ancillary programs in known versions). The LiveCD approach also makes installation pretty much a non-issue.

But I think it says something that a LiveCD can be the simplest solution to distributing an application....

Well

[Ranger]

If Microsoft says open source software is too complex, it must be so. Gosh, I'd have to say Microsoft nails it on the head every time they open their mouths about open source. This proves yet again that a Slashdot post has a intelligent headline and insightful blurb. Oh, and by they way, monkeys might flight out my ass.

He's kidding, right?

[1cebird]

Is Windows not too complex?

Open source too complex? Look at windows itself!!!

[Uzull]

...This is marketing bluff. I'm doing administation for both Linux and Windows (NT/2000/XP). At some point I really prefer to have a Linux system. Sometimes difficult to understand. But a Distribution from SuSE or Redhat is coming with nearly all features included and working. You know where the stuff is! When it comes to Windows administration, it is not only DLL Hell, it is also tools hell! Want to do something special, it always comes to a special tool you need to pay for...

pet peeve

[Anonymous Coward]

One pet peeve of mine is this: Red Hat uses their own scheme for version numbering standard libraries, so commercial software will frequently only run on Red Hat. The situation is approaching Microsoft DLL hell. I wish they would use the version numbers from the original authors, like every other distribution.

TFA comes from a different universe.

[marcello_dl]

How to export messages from outlook or migrate them to other accounts? M$ answer, get exchange or some shareware to do it (with free shareware screens and reminders, cool). Linux answer: backup one or two in evolution (WTF, evo guys?) and have them in an open and parseable format. Who's complex?

Eh?-Opposites distract.

[Anonymous Coward]

"Sorry, but I read this as "Choice is confusing - stick with what you are comfortable with. Hey look, that's us!""

Sounds like the KDE vs. Gnome argument.

"This sort of gibberish is what you would expect from the most popular product in the market who are being challenged for the first time in a while."

KDE.

if you decide

[Anonymous Coward]

and build linux from scratch, all of it is amazingly complex and quite a hurdle.

But i won't understand your argument while looking at distros like ubuntu ...

it is an issue of your workflow and design of it if you get hurdles of "complexity" in linux - you may review your process if you may be doing some wrong assumptions about your setup

Is Open Source too complex ?

[ookaze]

Why is this a problem or even a downside to MS ?

People who make Open Source never ask such question, so perhaps MS should make their engineers think of why this is so, instead of asking others. They are the company with lots of great minds, aren't they ?
Is this question even relevant at all ?

So what's new ?

[Anonymous Coward]

Open Source gives you freedom to innovate. Closed Source gives you lower maintenance cost. (But not many people have contractual maintenance either for Windows or Linux. They plan to piggy-back on the efforts of others, or give it their best shot themselves)
Eventually the closed source does not conform with the requirements, and you throw it away and start again.
With increasing supplies of Internet bandwidth, we find that Windows catches worms and viruses, Office documents drop keyloggers, old Office documents are no longer accessible, and people assemble POSIX-compliant OSes (Linux) and ISO-compliant office suites (OpenOffice.org) becuase they want them for their own purposes. The past is dead, long live the future.
Microsoft just need to lose interest in Windows and Office, and pick up interest in XBox, and the cycle completes.

Re:Open source is easier, if you know how to work

[DCGregoryA]

Microsoft is notorious for horribly misleading and vague error messages. Anyone whose dealt with Sharepoint's "Catastrophic error." message (and that's all it tells you, might I add, with no event log details or anything) can attest to that.

That being said, despite being a .Net programmer I tend to find open source pretty intuitive and straight forward. I could make the argument that lower level languages like C++/C and things like Php/Perl scripts are more complicated and granular than would be convenient for me, but then .Net's Linux version (Mono) and Java/JSP satisfy that dilemna.

Its more of a matter of a misconception due to people having limited experience in a non-Windows environment. Support is the big issue people have with open source software, typically, and accountability. I don't think complexity is a huge problem, though I will say its probably slightly harder to find experienced Red Hat admins compared to NT/2k admins. Marginally at least.

The man is full of crap

[Lumpy]

Open source is ready and IS used in business today. Many are even 100% open source because thay made that decision when they were small enough to not have the mess of an IT infrastructure that makes it near impossible to change over.

Where I am now we are 100% Open source except for vendor specific tools that are given to us by the vendor. The IT team here works hard to make them work under Wine so that we are 100% functional. New Sales people get over no windows and no Office2006 within 4 days and are as productive in open office and ubuntu as they were in windows.

Upper management and unskilled IT that cant handle standing outside their box are the #1 reason that open source is ignored and they buy yet another "solution" from a vendor.

REality - closed source vendors DO NOT give better support than Open source. Been there done that hearing the "that will be fixed in the next major release in 2 years" so many time I want to strangle them on the other end. MSFT tech support is 100% worthless from the OS level to the enterprise level apps (sql2003 enterprise)

I get better support from the people that write the Open source stuff. IF you PAY THEM the developers will bend over backwards for you.

The article is 100% fud.

Quirks, then you say Windows ..

[Blu-Ray]

working with windows clients since 3.x ..

quirks is what windows is most famous for, with every new windows version I had to learn another hundredplus or so ways to go around entirely new awkward behaviour..

And no way ofcourse to look into the source! and no chance somebody smart would find and fix that troublesome thing for ya ..

Re:If Open Source software is so complex to use...

[cecille]

In that case, why does slashdot spend so much time basing MS? why not just let OS software speak for itself?

nice typo

[Anonymous Coward]

Jason Pillai writes to tell us ZDNet is reporting that at last month's Microsoft Worldwide Parter Conference in Boston Ryan Gavin, director of platform strategy, claimed that one of the big downsides to open source is complexity

Worldwide Parter (sic!) Conference? A conference for those who want to part from Microsoft? Why doesn't Slashdot announce this earlier...

Oranges and Chimpazees

[forsetti]

The comparison is not Windows vs. Linux. It is Windows vs. Redhat Linux vs. Solaris vs. Novell Linux vs. Debian Linux vs. AIX ...

"Windows" is not comparable to "Linux" !!! One does not run "Linux" (generally), one runs "Redhat Linux", or "SuSE Linux", or "Debian Linux".

Therefore, it is necessary to compare the complexity of Windows with the complexity of a single Linux distro. If your Microsoft-friendly organization would be willing to consolidate on Windows, then your analogous Linux-friendly organization would be willing to consolidate on a single Linux distro, avoiding the multi-distro complexity.

If your organization is heterogeneous, and that is your "complexity" concern, then Windows is actually your problem! If you run or write software for multiple Linux distros, AIX, Solaris, HP-UX, and Windows, which one is the odd-ball? Admin'ing *nix systems is all similar, while Windows is *much* different. Porting from one *nix to another is easy, compared to porting *nix to Windows or Windows to nix.

I will grant that MS Windows provides a tighter integration for MS SQL, Exchange, and AD, than (for example) Novel or Redhat Linux provides for the database, groupware, and authN/authZ software included in their repositories. That is the cost of flexibility, which is oft balanced by the savings of flexibility. But this article attempts to pin the cost on the variation of distros, which is not correct.

The comparison is not Windows vs. Linux. It is Windows vs. Redhat Linux vs. Solaris vs. Novell Linux vs. Debian Linux vs. AIX ...

Re:Oranges and Chimpazees

[stubear]

He's not talking about the complexity compared to Windows, he's talking about the complexity of developing for a shifting environment. Linux comes in hundreds of different dstros, each with varying degrees of customization, not to mention the ability to completely customize the source itself, further adding to the target that vendors would be aiming for.

Re:Oranges and Chimpazees

[forsetti]

Exactly! Why would a vendor be willing to limit themselves to write code only for Windows, but not willing to limit themselves to a finite number of (or even a single) Linux distribution? The analog to a Windows-only software vendor is a Redhat-only software vendor or a AIX-only software vendor. Redhat is not a shifting environment (at least, no more so than Windows). There is no analog to a "Linux" software vendor in the Microsoft world, as there is no analog to Linux in the Microsoft world (again, the analogy is "*Redhat* Linux" or "*Novell* Linux" to Windows).

Vendors do not generally set a target of "Linux". Vendors set targets like "Supported on Redhat Enterprise 3 and SuSE Enterprise 9", etc, based on a finite number of supported configurations, much like saying "Supported on Windows 2000, Windows 2003 Server, Windows XP Professional, Vista Professional, Vista Ultimate, Vista Home, Vista Media ..."

As an aside, I generally avoid statements like "Linux comes in hundreds of different distros...", and refer instead to "Redhat Linux" and "Novell Linux" as completely distinct operating environments, with distinct support, configuration, and software requirements.

Unedited version

[Anonymous Coward]

IBS Synergy had started developing products for the Linux platform back in 1998 but gave Linux the boot in early 2004, and now builds its software on the Windows platform. Lim said this was because the company's developers were spending more time surfing slashdot; er, I mean, "hunting for Linux technical support on the Web", and had less time to focus on actual development work.

Re:Open source is easier, if you know how to work

[qray]

Interesting, I've never had need to talk to a MS-SQL or any database developer to use any of the database products I've used. I've worked on some pretty complex databases and queries as well.

My experience with PostgreSQL some time back was much worse than any experience I had with SQL Server. The online documentation failed to match anything resembling the actually installation that existed. After having expending a fair amount of time figuring out my Linux install and how to build the kernel hitting this with Postgre SQL put me over the edge. This was a side project I was working on and I just didn't have the time to go searching across the Net to figure the mess out. This was sometime back, hopefully things have changed.

Having contributed to open source, what I found is a lack of coherent design. Often, it seems, the way things get done is for some "smart" person to come along and take on something and complete it. Unfortunatley usually these "smart" people don't enjoy creating designs and documentation so after they move on to the next fun project, the lesser people poke and prod what they have created trying to add features and fix issues but never really understanding it. Eventually it gets to the point where another "smart" person comes along and rewrites it because it's become a mess and no one really understands it.

The fit and finish often seems lacking in open source. I tried to convert my wife to Open Office. It just didn't work. The spreadsheet just had too many issues for what we were doing. I ended up reinstalling MS Office. I think some of this may be due to the lack of coherent design of the system as a whole, but I'm not familiar enough with Open Office to make a call on it.
--
Q

Re:Open source is easier, if you know how to work

[orbitor]

Getting started. Period. Microsoft has always been very good at getting a new user up and running with their software. And as long as the user stayed within the perdescribed road map, they would probably be in decent shape. Once the user had to do real work with the product, however, they were fucked. But, by then, the user was already hooked. Anyone who has worked corporate knows this story. Your team and your management have a dozen or so meetings about which tools will be applied for a certain project; the asshole in the team who is an idiot, but has management's ear, begins their pontification on the wonderous joys of using the latest version of MS PRODUCT because he has a beta copy of it and lookie how easy it is to do this TRIVIAL TASK. At which point management claps their hands, grunts and exchange nods of approval. The PO is signed and everyone on the team gets a copy of MS PRODUCT. Sure, not everyone can install it right away because not everyone is at the same patch level as the alpha dork who recommended the crap (and who is now, by the way, taking training classes on THE NEXT BIG THING. And why not? His job here is done, he got you into this mess). So, you struggle with it for a while, finally get something simple to work by following the tutorials, start to feel pretty good about the situation, confer with your colleagues that "Hey, maybe this isn't so bad." Then, management comes to you with the requirements document and you immediately realize that you've been had. But, it's too late.

Contrast that with open source. Sure, there are a great number of projects out there that have great documentation, and some do have tutorials. Wonderful. Those are the projects that succeed (ie. Apache, MySQL, PHP, etc). But, let's face it, those are in the minority. How many projects consist of only source code? That's not to knock the developers; but, honestly, documentation is boring work and by the time it's done, it's already out of date. It takes a great deal of time to keep the implementation in sync with the documentation. If the developer is in it for the glory (how world reknown would Torvalds be if he would have wrote great documentation about Minix back in '91 and released it as open source?), the code is what matters, not informing the user on how to get started with the software.

As for the IIRC comment, sure that works when the project is large enough that there is a (and I hate using this word) community built up around it where there are users knowledgable enough to answer your questions that make themselves available and take the time to answer questions. Of course, the other side of the coin is the poor newbie who gets flamed to shit the first time they ask a question. The common response is RTFM, dumbass! Which I can not disagree with if there is an FM to begin with. I have learned over the years that no matter how unpalatable documentation is, it perfectly serves its purpose when the user of your software does not contact you when they are installing it, learning it or pushing it to an extreme, making it do shit you hadn't thought of; rather, they contact you to show off what they did or to inform you of a bug that they have found and they are forwarding their idea for the fix.

Again, really, it comes down to the percieved start up cost (as in time and effort, not money). Microsoft has always been very good at that part of the game. There are a huge number of open source efforts that could be more relevant if they learned this lesson.

- orbitor

It's NOT complex

[ajs318]

It's NOT complex to develop software to run on an Open Source platform. It only becomes complex when you try to conceal the source code from users. The configure - make - install process generally works well, unless the packager ballsed up and forgot a dependency that didn't get autodetected.

This method worked really well for Microsoft Windows -- absolutely nobody can ever copy it because they can't see the source code. Wheras all the high-end proprietary Unix programs that came in a source tarball without permission to redistribute, well, the world and his cat have copies of all of them.

We need either a law mandating user access to source code, or a decompiler good enough almost to guarantee this anyway. Then, the only thing that makes it hard to develop on Open Source platforms will not be there. And, as a side effect, we'd all be in more in control of our computers.

Re:It's NOT complex

[ErikZ]

I've tried to install TA Spring on Debian and at this point I'm of the opinion that there shouldn't be libraries on the end users computer at all. Put the onus on the developer to decide if a library is too buggy. Make the source code available, but ultimately, just have binaries. How many years have we been playing with package managers?

Keeping track of libraries and versions is just crazy and offers no benifit to a user who just wants to install the software and get it to work. Yes, it will increase the download size of the software. Big deal. It's a far better option than not being able to get the software to compile.

Sorry, buddy, wrong cause identified

[meburke]

I keep going over this with people all the time: Complexity is a by-product of design decisons, including features. The more connections you make, the more complex the system.

I believe that attributing complexity to "Open Source" is a mis-applied cause. A cause is defined as something both necessary and sufficient to explain the effect. The article does not describe the problem domain in enough detail to support the argument. The complexity is a result of arithmetic, not "Open Source".

In the old days, doing assembly language, I would build my systems using Decision Tables. The complexity is directly related to the number of decisions (5 rules = 2^5 = 32 potential tests, 10 rules = 2^10 = 1024 potential tests.) Athough all the tests are mathematically possible, a large number of test don't make logical sense. But the amount of complexity still increases rapidly with the number of decisions traversed by the system.

UNIX (and LINUX) have simplified the state of the system at the OS level by accessing everything as if it were a file. Complexity is added after that by the number of states and messages that an app has to keep track of in order to traverse the decision points correctly. Complex applications sometimes require additional state computation in order to work with other applications. So, as you work your way up from a simple text terminal program to a windowed environment, the complexity multiplies a LOT. Now, consider that there are a multitude of environments to run in, a multitude of desires and possibilities for applications, then if you want your application to run in more than one simple environment you have to make the installation choices more complex. (Therefore, I maintain that writing an application to run on both Gnome and KDE increases the complexity almost as annoying as writing it for two different operating systems.) Then add one more choice: How do you keep the app updated? (rpm? apt? source? etc.?)

Dang! This whole post is just another way of saying that the author didn't distiguish between "complexity", "ease-of-use" and "ease-of-administration" and blamed the confusion on "Open Source".

My wish, of course, is that more designers of systems and apps would put more thought into resolving the administrator's headaches instead of just adding cool new features. My standard is to "write it to a directory" and be done with it, but I usually write stand-alone apps that don't require so many complex interactions.

Having said all that, I also want to say I'm grateful to designers like the FireFox folks who spend a lot of time making the app easily installable and self-updating.

s Open Source too Complex?

[ChronoFish]

No.

-CF

BS

[pedestrian crossing]

I'm all for freedom, but I find it silly that different distros keep configuration files in different locations, use different init scripts, use different install methods, have varying level of compliance with the LSB, are focused on either Gnome or KDE predominately, etc.

It does make it more difficult for a large company to develop for a Linux crowd in general.

Only if they want to develop proprietary software for Linux.

If they provide the source, then whoever maintains the distro is the only one who has to worry about issues that you are fretting over.

That's the whole purpose of distros.

I'm all for freedom

I would argue that you are not. Otherwise, you wouldn't be rolling out this old canard...

Re:BS

[CTho9305]

Only if they want to develop proprietary software for Linux.

If they provide the source, then whoever maintains the distro is the only one who has to worry about issues that you are fretting over.

That's the whole purpose of distros.

Even for OSS, that's just not the same as being able to distribute one package that works everywhere. On Windows (9x, 2k, XP Home, XP Pro, Vista's 7 versions), I can ship one binary package that works for everybody. Microsoft doesn't have to approve my package before making it easily available to users - any Windows user can download my one simple installer and have it work for them regardless of their version. Now look at Linux: there are many different distros, with many different package formats. I'd have to provide RPMs, DEBs, tarballs, and probably multiple versions in each format (since it might depend on different packages for different distros). Users would have to know which package to download. If the experience is going to be easy, I have to beg the distro's maintainer to provide an official package--some distros are very slow to add new products.

A real-world example of this is SeaMonkey. How long will it be until Debian users can install the software easily? Windows users can have the latest version as soon as we ship it. Linux users generally have to wait for their distro to provide an updated package. We do provide tarballs which you can extract anywhere, but that's not really user-friendly... we also provide an installer, but then you're using a method of adding software other than your distro's standard package management. We do ship source, and anyone is welcome to build it and include it in distros, but the vast majority of people just want to install a binary using whatever method they normally use (e.g. google for the website, download an installer, or search Synaptic, etc).

Re:BS

[Anonymous Coward]

Only if they want to develop proprietary software for Linux.

If they provide the source, then whoever maintains the distro is the only one who has to worry about issues that you are fretting over.

That's the whole purpose of distros.


This is the sort of attitude that results in a fragmented, divided, and incompatible linux universe. The source isn't always the answer. Just because you have source doesn't translate it to being easily cross compatible with every distro (heck, it's oftentimes not even cross compatible with different versions of the same distro).
The linux mentality of having to reinvent the wheel every six months is what keeps it from being a serious contender in the Desktop market. Unless one distro becomes exceedingly predominant (say Ubuntu)Linux will never seriously challenge Windows or even Mac OS X. Linux succeeds well in the server market because the majority of servers run LAMP or LAM(x). It also succeeds in the embedded market because the kernel is highly portable and, again, the applications are limited. However, if you want expansive applications that go beyond LAMP or embedded and want to seriously enter into the end user market, then you will need a serious committment towards uniformity in the distro.
As for proprietary software, remember that the computer industry as you know it was a result of companies being able to sell software in the first place. Without proprietary software we would never have had the computer revolution in the first place.
Limited choice is good, but too much choice is detrimental.

Re:BS

[rainman_bc]

Even for OSS, that's just not the same as being able to distribute one package that works everywhere. On Windows (9x, 2k, XP Home, XP Pro, Vista's 7 versions), I can ship one binary package that works for everybody.

Depends. Do all 9x boxen have the .net runtime? Do they all use the same MFC? Only the most basic program can you make it run on any platform. Or are you coding in Java???

Microsoft doesn't have to approve my package before making it easily available to users - any Windows user can download my one simple installer and have it work for them regardless of their version.

Odd, I use FC5, and I use third party yum repositories for any software not officially provided by the main Fedora repository.

Now look at Linux: there are many different distros, with many different package formats. I'd have to provide RPMs, DEBs, tarballs, and probably multiple versions in each format (since it might depend on different packages for different distros). Users would have to know which package to download.

That's what apt-get or yum are for. And with synaptic or yumex it's a piece of cake.

If the experience is going to be easy, I have to beg the distro's maintainer to provide an official package--some distros are very slow to add new products.

Again, see my comment above about third party repositories.

A real-world example of this is SeaMonkey [mozilla.org]. How long will it be until Debian users can install the software easily? Windows users can have the latest version as soon as we ship it. Linux users generally have to wait for their distro to provide an updated package.

That's odd. pbone.net has Seamonkey in their repository. If I want it, I can get it here ...snip

but the vast majority of people just want to install a binary using whatever method they normally use (e.g. google for the website, download an installer, or search Synaptic, etc).

See comments above about third party repositories. If you want to be bleeding edge, that's your problem, not the distro's.

Re:BS

[Enderandrew]

Thank you. That was exactly what I'm getting at.

I love OSS. I love GNU/Linux. But I'd love to see stronger standards, and distros complying with them even more.

Re:BS

[bcrowell]

You have a good point, but I'm not convinced that it completely addresses all the possibilities.

Let's imagine, for example, that you're marketing an MMORPG for Linux, and let's imagine that the software is free and open-source, and you're just charging customers per hour to use your server. You're ready for the big roll-out, so what do you do? Post on a Debian mailing list, ask for someone to package it for Debian, and then wait to see if someone gets around to it? And then do you do the same for every other distro?

As another example, I've heard people who write open-source Linux drivers for printers, scanners, etc., complain that the kernel changes too frequently, and it's too much work to keep fixing their driver when a new kernel release breaks it.

Here's another example, and this one is real. I wrote a piece of open-source software that uses the Perl/Tk GUI library. Turns out that some distros (including ubuntu) ship with a version of GTK+ that interacts badly with Perl/Tk: when you run a GTK+ app and a Perl/Tk app side by side, the GTK+ app broadcasts X Windows events that cause the Perl/Tk app to crash. I submitted a patch to Perl/Tk (it's a bug involving dereferencing a null pointer), but the maintainer of Perl/Tk says on his web page that he no longer has time to do much work on it; roughly a year has gone by, and he has apparently never applied the patch. For me, it's a major issue in my life, but it would be if you were an application developer who was actually trying to make money off of open-source -- that is supposed to be possible, isn't it?

Re:BS

[bcrowell]

Sorry, the last sentence should read "For me, it's not a major issue in my life..."

Re:BS

[Frumious Wombat]

Except the issue is that if the software doesn't work as expected, after the distro maintainer finally gets around to packaging it, the customer is going to call *you*, not *them*, to complain. Even if the package is free, someone is going to call (email, IRC, telepath, whatever) you and complain that they're running UltraLinux 3.14, and your package isn't built correctly, like it is for RedHat users. They'll never write the distributor and complain, "you put all of the data files where the program can't find them".

If you're a business, I can see wanting to have finer control over installation, and that yielding the decision to not support certain distros, or not develop for Linux at all. This goes to show that the GNU Zealots are right; Linux is just a kernel, onto which many, many, operating systems have been layered.

Re:BS

[Anonymous Coward]

I would argue that you are not. Otherwise, you wouldn't be rolling out this old canard...

If this is "freedom" and I want to run proprietary software, why are you complaining? Is it "freedom" only if I do what you think is right?

Re:BS

[CTho9305]

Depends. Do all 9x boxen have the .net runtime? Do they all use the same MFC? Only the most basic program can you make it run on any platform. Or are you coding in Java???
That's BS. SeaMonkey (and Firefox, and Thunderbird, and thousands of other appliactions) isn't a .net application or trivial MFC application, yet somehow runs on all versions of Windows from the past decade (even NT 3.51 with some minor caveats!). It's also not written in Java. It's a pretty serious piece of software.

Odd, I use FC5, and I use third party yum repositories for any software not officially provided by the main Fedora repository.
[snip]
See comments above about third party repositories. If you want to be bleeding edge, that's your problem, not the distro's.

So basically what you're saying is, in order to get an equally good experience as a user when your friend says, "Hey, there's this new SeaMonkey project that continues the Mozilla suite and adds some cool features", you have to hunt for third-party repositories and set them up in your package manager? Explain to me how this is not more complicated than Windows where you just go to the program's site and download the installer.

Re:BS

[Dadoo]

On Windows (9x, 2k, XP Home, XP Pro, Vista's 7 versions), I can ship one binary package that works for everybody.

You're joking, right? If I had a dollar for every time a software vendor told me "Oh, we don't support Windows XP", or at the very least, had to be run under the Administrator account, I could retire.

For popular OSS packages, the most difficult thing you need to do is "./configure; make; make install". That's a lot easier than even messing around with some of the licensing procedures for closed-source software.

Re:BS

[Anonymous Coward]

You're sort of making his point, by showing that for FC5, the latest version available is 1.0.2, while the latest release is 1.0.4 - you're not just slow to get new features, you're behind on security fixes.

Re:BS

[rainman_bc]

That's BS. SeaMonkey (and Firefox, and Thunderbird, and thousands of other appliactions) isn't a .net application or trivial MFC application, yet somehow runs on all versions of Windows from the past decade (even NT 3.51 with some minor caveats!). It's also not written in Java. It's a pretty serious piece of software.

That's because the UI is written in XUL, and the installer includes all the libraries for it, for all win32 platforms. Not all software for windows is so lucky and portable. You're mistaken about windows apps being portable.

Explain to me how this is not more complicated than Windows where you just go to the program's site and download the installer.

Seriously, have you ever looked at the seamonkey download page where they offer a Fedora Core rpm build?

If you want the ability to be able to download and install, just use Fedora. Don't bitch if you're using a geek Linux like Ubuntu or Gentoo or Slackware or something. Fedora is a pretty usable Desktop OS right now, and most download sites offer rpm builds of their software. Not everything needs to be in a yum repository.

Look, there are other problems with desktop Linux, but the ones you cite aren't the bad ones. Hardware compatibility for example. Or kernel modules you need to recompile every single bloody kernel update. Drivers that only work under ndiswrapper. WPA support in Linux wireless-tools package instead of wpa_supplicant. Those are issues. Wireless support should be friggin' perfect by now, yet lags behind the windows world by a long shot. Those are barriers to penetration much more than the availability of a binary installer (which btw some companies also offer).

Re:BS

[labratuk]

As a software developer you should not make packages. That is not your job.

A distribution's job is to make sure a whole pile of software (nearing 10,000 packages for debian) works together in a consistent and reliable way. This is one of the reasons why freenix systems are so reliable. The resultant system is not a hodge-podge of different software authors' wills trying to bend the system to their personal tastes.

So, unless you know every distribution better than the distribution's developers, you shouldn't build your own packages. That's not to say you shouldn't be helpful to packagers.

Re:BS

[llefler]

Depends. Do all 9x boxen have the .net runtime? Do they all use the same MFC? Only the most basic program can you make it run on any platform. Or are you coding in Java???

Surely you aren't serious. I have written enterprise n-tier applications that have run on Windows 9x, NT, 2000, XP and 2003. I don't care about .NET or MFC. I have written Windows services that run (same exe) on NT, 2000, XP and 2003. And no, I don't do Java either. I also don't have to worry about what version of kernel I'm running or if I have the right common controls library.

Re:BS

[Rich0]

I'm not intimately familiar with Fedora, but many distros backport security patches. So, even though you might be two versions behind bleeding-edge you may in fact be completely secure. Backported patches have the advantage of being less likely to break things.

Re:BS

[CTho9305]

That's because the UI is written in XUL, and the installer includes all the libraries for it, for all win32 platforms. Not all software for windows is so lucky and portable. You're mistaken about windows apps being portable.

You're full of it. If you've ever read Raymond Chen's blog you'd know just how much work Microsoft puts into maintaining binary compatibility - and the results show. Maybe you're the one person in a million who uses the few applications that do break, but in my experience, when upgrading to new Windows versions, only a tiny number of programs have problems (mostly games, which tend to abuse low-level features... Raymond's blog provides numerous examples where the applications that do break tend to be misusing APIs).

Getting back to SeaMonkey, it's not like the whole backend is reimplemented for each Windows version. The vast majority of the code really isn't at all affected by the Windows version. There's a lot of C++ code in SeaMonkey too (it's nowhere near being just XUL + JS on top of Gecko), and of all the SeaMonkey and Gecko C++ I've written I've only had to worry about the Windows version once - when I was modifying splash screen code to take advantage of new Windows 2000 features. There are certainly other cases where the code is optimized and takes advantage of features of later Windows version, but it's rarely more than a tiny change to accomodate all versions of Windows. I don't recall ever worrying about the Windows version in the rest of my patches at all.

Seriously, have you ever looked at the seamonkey download page where they offer a Fedora Core rpm build?

If you want the ability to be able to download and install, just use Fedora. Don't bitch if you're using a geek Linux like Ubuntu or Gentoo or Slackware or something. Fedora is a pretty usable Desktop OS right now, and most download sites offer rpm builds of their software. Not everything needs to be in a yum repository.

So if I want even non-proprietary applications to be readily available, I have to use one of a tiny number of distros (ignoring the fact that I'm sure there's some other project that provides DEBs but not RPMs). That seems to me like you actually don't have any real choices in the Linux world.

Look, there are other problems with desktop Linux, but the ones you cite aren't the bad ones. Hardware compatibility for example. Or kernel modules you need to recompile every single bloody kernel update. Drivers that only work under ndiswrapper. WPA support in Linux wireless-tools package instead of wpa_supplicant. Those are issues. Wireless support should be friggin' perfect by now, yet lags behind the windows world by a long shot. Those are barriers to penetration much more than the availability of a binary installer (which btw some companies also offer).

Those are certainly larger barriers to adoption, but I do think that the more limited binary compatibility with Linux is not something that should be ignored.

Re:BS

[10101001 10101001]

I'm sorry, but is the discussion that the delivery of products is more complex or that the use of products is more complex under Linux?

If it's the former, I'd whole heartedly agree. That's a design descision made by distros to limit the malicious and/or non-malicious damage caused by allowing every piece of software into the main distro line. People can still choose to walk outside of it, and distros surely are constantly undermanned to do the job of adding legitimate programs at an optimal rate. But, distros are focused more on providing for the users, not the distributors. After all, distros *are* distributors, and they damn well recognize they're doing the hard work of distributing.

If you're more interested in having an easy and complex life than helping your users, don't be surprised if the Linux approach seems "unfair". It's pretty unfair when developers only release for Windows or the PS2 or whatever platform. Users have learned to cope. Targetting Linux (and the *BSDs) might just mean requiring the developer to cope.

Oh, and as for your example:

>A real-world example of this is SeaMonkey. How long will it be until Debian users can install the software easily? Windows users can have the latest version as soon as we ship it. Linux users generally have to wait for their distro to provide an updated package.

That last sentence nails the answer. Debian isn't focused on being up-to-the-minute. It's focused on stability. Bitching that Debian doesn't update often enough is effectively bitching at the users for not wanting to update often enough. That's not exactly a great focus on user wants.

Re:BS

[jrobinson5]

Yes, but at least source makes it physically possible for someone, somewhere to (legally/practically) port it to a different platform.

Re:BS

[Identifiable Coward]

Those are certainly larger barriers to adoption, but I do think that the more limited binary compatibility with Linux is not something that should be ignored.

You seem to be forgetting that Linux runs on much more than just x86. I like the fact that I can use the same software (I run Gentoo) on my SPARC, my x86-32 and x86-64, as well as my PPC systems. Now show me how to do that with a binary installer.

Re:BS

[Identifiable Coward]

Except the mere fact that you mention the "right common controls library" means that you have at one time or another run into that problem. Last I recall it was a problem on Windows. I've never heard of a "common controls library" for Linux.

Re:BS

[llefler]

Actually, no, I haven't. I have seen developers years ago that distributed the DLL with their installs because they needed a specific version or had made customizations. It was rare, and they caused their own heartaches. The Linux equivalent would be needing the right widgets. Such as needing a specific version of QT or GTK.

I'll have to run an experiment this weekend. A couple weeks ago I got a new XP machine at work and installed Delphi 7 on it. Prior to that I was using W2k. Applications that I built on my old machine compile clean on my new machine with no modifications. I also have a Debian box that I recently built for MythTV. I'll see if the Kylix that came in the same box with D7 will run with a current version of QT. (Kylix was written to support QT 2) Want to take bets on it working with a clean, current QT install?

Parting

[6031769]

FTFS: Microsoft Worldwide Parter Conference

Is it clear now that this is a conference for how MS is going to part fools and their money?

Complexity Level

[MrCopilot]

I don't know about the rest of you guys but, I kind of like MS thinking Open Source is too complex for them.

To make tasks appear simple usually requires a very complex but well thought out plan, implemented with some elegant code.

So I submit maybe we still aren't complex enough. I like the way my brain hurts at the end of every day, makes me feel alive.

Re:Open source is easier, if you know how to work

[Gr8Apes]

This is 100% true - life is too short. If a MS product doesn't work in 15 minutes, delete it. Whoops, there's a large disturbance in Redmond as 200 million windows systems just converted to something else.

Seriously, though, if you'd like to know a serious point of weirdness (and where MS yet again failed to support a standard) is the X.500 adaptation in Exchange up to at least 5.5. Organization (the "O" in the addressing) is supposed to be case sensitive. In Exchange up to at least 5.5, it isn't. You can connect organizations with "Org", "orG" and "ORG" together. This does, however, cause some issues with messaging that does honor the X.500 spec.

Oh, and you cannot change the organization either - because it's populated throughout the "standard" JET DB that underlies Exchange.

Long story short - 1 typo screwed up a system years after implementation, because it wasn't caught until a configuration change was to be made.

To address the MS SQL DB item - try clustering the POS. Oracle, IBM DB/2, PostgreSQL or MySQL anyone? MS SQL is a nice replacement for Access, as it actually works. For enterprise use, however, get a real DB.

Two words....

[moosesocks]

For my sins I have used a lost of operating systems over the years and they all have their pros and cons, the one thing that seems common across them is that the more scarey they look the less likely they are to break because people don't mess with the difficult ones. Most failures are caused by human error (it's just that no one admits to it) and making server OS's look familiar tempts people to fiddle.

SCO OpenServer.

Scary, old, unsupported (difficult to find anyone willing to work with it), and extremely easy to break, even by end users.

Then again, this is probably the result of bad karma for still running a serial-line network in 2006.

To draw another analogy, Token Ring is also scary, and very easy to break. On the flipside, Mac OS X is much "simpler" to an end-user than Windows is(and to a certain extent, to the developer as well) -- it is also much more difficult to break. I've never seen an OS X installation trashed in such a way that it couldn't be fixed by creating a new user profile. Granted, this is due to OSX's UNIX underpinnings, but the fact remains that it's pretty undeniable that it's a simpler system to use for the user.

Re:Two words....

[jgmitchell]

On the flipside, Mac OS X is much "simpler" to an end-user than Windows is(and to a certain extent, to the developer as well) -- it is also much more difficult to break.

How is it simpler? It all depends on what a user is familiar with. I find it simpler to type cd /home/user/docs/dir/stuff than to double click a bunch of folders to open some file. But then again I use those commands everyday.

Re:Two words....

[jacksonj04]

What gets me about OS X is that not only is it easy and solid for end users, it's also easy and solid for home admins. I've seen people with no previous experience set up a whole family with accounts (and appropriate settings) on a Mac simply by following the prompts. And they still couldn't break it.

Windows and some Linux flavours (Ubuntu has a nice one) have cottoned on to multi-user in the home, but still is a bit wobbly on the admin side. Vista's new user admin seems to be almost OS X in simplicity (With scary user tracking options), so I'll have to see how that works in the field.

Re:Two words....

[Kuxman]

I (like you), also have my file system memorized. I know where everything is located, and am a sorting-nazi, so "cd /home/user/docs/dir/stuff" is the way I navigate. But Look at most people's filing cabinets, and it's a mess. Look at their computer.... also a mess. For me, if I'm in unknown territory, it's quicker to click through a bunch of folders rather than cd folder ... cd another_folder... etc.

Re:Two words....

[severoon]

I'm always stunned by the group of people that don't recognize the value of proprietary software packaging as a benefit to business. Some of the proprietary stuff sucks, but clearly, some of it doesn't. One of the maxims of business is, has, and will always be that the customer is always right. Why can't OSS defenders see this? If a business complains or fears that OSS will be too complex, then, whether it's too complex or not, IT IS TOO COMPLEX. It's either an image problem or a usability problem, either way it's a real, live problem.

One of the things I hate about the OSS community is the I-know-better attitude they take. You don't know better, because you don't know the business of the customer as well as they do, and you probably don't know what they're expecting from a given package as well as they do either. This stuff requires a lot of work. I'm a big proponent of OSS, and I think that someday it's going to take over the world and be the primary way that software is written, but I'm always frustrated when topics like this come up and I'm reminded of how much work there is left to do. From users to developers, it's sooooo hard to contribute to OSS popularity (either as user or developer) that many of the best advocates on both sides are left behind.

As far as Macs go, for the first time in a long time I used OS X a couple of weeks ago. I was trying to help a friend use their all-in-one printer to scan in an image. My conclusion: OS X sucks...it's horrible and clumsy to use. Every time I say that, Mac defenders always tell me the same thing: it's not that it's bad or good, it's what you're used to, and you're not used to the UI.

Well, ok, let's look at that argument. Why does it take me a click and a drag and another click-drag to size a window where I want it? Every other window manager I've ever used (and I've used probably more than a dozen) if any one corner of a window is where you want it, you drag the opposite corner to where you want it and you're done. Not so...Mac decided that something so simple should cause a new user to hunt around for a few minutes trying to figure out why the window doesn't respond in the expected way. Ah ha! It's only the lower right corner that can be moved. But then that means if you want to drag the upper left, you have to resize the window using the lower right...but then you hit the bottom right corner of the screen and it's still not big enough. So the actions are ordered as well...*first* you have to drag the upper left where you want it, and the drag the lower right back where it was.

Let's talk about iLife...so back to scanning my picture. I put it in the scanner of the all-in-one and hit the scan button, it pops up a menu asking me what I'm scanning (do I want OCR, bitmap, etc?). Great, so I hit image and it scans it. On the desktop, an app automatically pops up a gallery of images and another small window with a thumbnail of the image. There's a note under the thumbnail telling me to drag and drop it into the gallery application. Oh great! Ok, this is really easy--maybe Macs are better.

So I create a new folder in the gallery app and drag the image to it. Upon releasing, however, the image doesn't show up in the gallery in that folder. Huh. So I try it again, as the thumbnail is still there giving me no indication that it actually went anywhere in the gallery app. After much frustration, I figure out that there's a magic default folder (I forget the name even now) where that application has decided all new images go before you can sort them into other folders. As I dragged it to multiple locations multiple times, I created several copies of the image in that default location. Once I discover this and enter the default location, I notice that the user who I'm helping finds this as tedious as I do--because you are forced to organize and copy in two separate steps, he hasn't gotten around to organizing a few hundred photos which just sit in the default location, never having been sorted (and not like to be).

After hunting through that

Re:Two words....

[ArwynH]

It's not so much having your fs memorised as Tap-Tap being faster than Click-Click. At least for me it is.

My home dir is a mess and I still find it easier to cd docs/[tab][tab]colour[tab]/[tab][tab]pi[tab]. Much faster and simpler than [click][scroll][click][scroll][click], even if it doesn't look it.

I can't remember from whom I heard that the true beauty of X was that it allowed multiple consoles to be on the screen at one time. Oh how right he was... :)

Re:Two words....

[Anonymous Coward]

> My home dir is a mess and I still find it easier to cd docs/[tab][tab]colour[tab]/[tab][tab]pi[tab]. Much faster and simpler than [click][scroll][click][scroll][click], even if it doesn't look it.

Then why not open a Finder window, and navigate using the keyboard? Command-down-arrow isn't any slower or more difficult than hitting tab repeatedly. :)

Re:Two words....

[styrotech]

One of the maxims of business is, has, and will always be that the customer is always right. Why can't OSS defenders see this?

They can - it's just that with most open source projects the end user isn't a customer. For the most part end users are just freeloaders who don't give anything back to the project. For an open source project, a few good contributors (eg developers, testers, writers, artists, donors etc) is worth far more than hundreds of end users that don't contribute. Contributors are effectively the 'customers' of an open source project not the end users. The project is the contributors.

Re:Two words....

[moosesocks]

Stop trolling.

You're right. It's all about what you're used to, and what you're capable of doing. However, you didn't actually provide any logical reasons to prove that OS X/Windows/Linux is superior/inferior.

There is an obvious visual cue that OS X windows are meant to be dragged from the bottom right corner present in every window that can be dragged. Many Windows applications behave the same way, and provide this same cue. I'd say that OS X is superior in this respect, simply because it's consistent across the board.

Comparing the manner in which your 3rd-party photo gallery app works doesn't really help compare the two either. If you were talking about iPhoto (I couldn't tell -- those paragraphs were completely incoherent), again, it's no basis for comparison, as Windows includes no similar application.

Not having to remove spyware is also a nice advantage for OS X (at the moment at least).

People really need to stop scrutinizing Apple down to a hair, whilst Microsoft's discovers a dozen gaping holes in its operating system every week, and Dell laptops burst into flames on a regular basis.

Re:Two words....

[severoon]

I couldn't disagree more. You can have the best application in the world. If no one is using it, it amounts exactly to a pile of doggie doo.

Re:Two words....

[severoon]

I can't understand why anyone would reply to a post they couldn't figure out (re: your "incoherent" comment).

In any case, let's assume you're right (you're not) about the "obvious" visual cue. Why do I need to make two separate click-drags if all I want to do is move the upper left corner of a window while leaving the lower right where it is?

I believe I was talking about iPhoto. Your response to my problems with it is nonsensical. The point of including default apps on the Mac is that it's supposed to be ready to go when I plug it in with no installations necessary. If the apps they include by default suck, how is that more usable than a PC? I submit that it's less usable--not only do I have to figure out how to install an app that works, I have to figure out how to uninstall an app that doesn't.

Not having to remove spyware??? I've never had to remove spyware from my home PC, or my home linux box...and my PC is actually under attack. I hate to think how the average Mac user would secure their systems against spyware if they ever gained more than 10% marketshare in the home and fell under attack in a similar way. Depending on the unpopularity of the platform seems to me to be a form of security through obscurity, which is hardly better than none at all.

You're right about one thing, though...MS blows. That's why I don't run any MS software -except- Windows on one of my two boxes, and the only reason I run that is so I can have Photoshop because, sadly, linux has nothing that can compete. (Please, don't bring up the gimp until you have actually used it to do significant post-processing akin to what one might do for showing fine art in a gallery.)

Re:Two words....

[turbidostato]

"If a business complains or fears that OSS will be too complex, then, whether it's too complex or not, IT IS TOO COMPLEX"

Even when the company that so says is Microsoft?

Well, next time I hear from Coca-Cola that Pepsi sucks (or the other way around) I'll know it's true because, you know, when a bussiness complains or fears that Coke sucks it means IT SUCKS.

Complexity....

[IckySplat]

I friend of mine once asked me what I liked about Linux...
I told him "I can make it do just about anything, with a bit of work"

Then he asked me what I didn't like about Linux, and I told him...
"To make it do anything, I need to do a bit of work"

Which I think is a pretty fair and honest account :)

Not too complex for Cyber Goats

[Anonymous Coward]

http://www.ubergeek.tv/article.php?pid=54

null and void.

[everphilski]

but then .Net's Linux version (Mono) and Java/JSP satisfy that dilemna.

Yeah, cause Mono is based on ECMA standard (generated by Microsoft) and Java isn't open source. And since C++ and PHP/Perl are "more complicated and granular than would be convenient for you" I guess your arguments are null and void. In your case, closed source is greater than open.

Re:null and void.

[DCGregoryA]

In your case, closed source is greater than open.

In my case, getting the job done in a way that satisfies our clients and earns us revenue is greater than caring about politics. Java's closed-sourceness or .Net's ECMA standards don't prevent them from being used widely to the benefit and profit of many corporations across the globe. Go figure.

Re:null and void.

[admdrew]

In my case, getting the job done in a way that satisfies our clients and earns us revenue is greater than caring about politics.

"Client satisfaction" and "higher revenue" don't always imply the project was coded well. Now, I'm not suggesting designing decisions should be based on technically ill-informed politics, but the "ends justifying the means" school of thought makes for messy software.

Java's closed-sourceness or .Net's ECMA standards don't prevent them from being used widely to the benefit and profit of many corporations across the globe. Go figure.

...so the "everyone else is doing it" mentality is always a good argument?

Re:null and void.

[DCGregoryA]

"Client satisfaction" and "higher revenue" don't always imply the project was coded well. Now, I'm not suggesting designing decisions should be based on technically ill-informed politics, but the "ends justifying the means" school of thought makes for messy software.

That's quite a leap to take from what I was saying (notably, the ends always justify the means). Additionally, maintainability is typically a factor in client satisfaction, and one not to be neglected. I'd like to know, precisely, how Java makes for messier solutions than C++.

Ironically, I'd take the exact opposite standpoint.

...so the "everyone else is doing it" mentality is always a good argument?

The topic isn't cocaine usage, we're talking about tried and proven technologies used in high stress environments. The poster whom I was responding to was indicating that "Java's closed-sourcedness and Microsoft's ECMA standards" preclude them from being viable choices, which is absurd. If you'd like to make a real argument against them, go for it.

Re:null and void.

[admdrew]

The poster whom I was responding to was indicating that "Java's closed-sourcedness and Microsoft's ECMA standards" preclude them from being viable choices, which is absurd.

Your indication that they (Java and MS's solutions) are viable only because they're used extensively is *also* absurd. Given that I'm a .NET programmer by trade, I certainly don't disagree with your choice of language... but there are other merits in favor of these two systems.

Re:null and void.

[DCGregoryA]

viable only because they're used extensively is *also* absurd.

Only viable because of that? Where did I say that? I'm not going to debate why Java/.Net is good because, frankly, I don't have to. Its a well established fact (which was my point). May as well start debating why breathing is good.

Re:null and void.

[everphilski]

You made the implication that Java and MONO were open source. MONO is, but it is based on Microsoft's code, and Java is anything but open source. I was just pointing out the flaw in your argument.

Re:null and void.

[DCGregoryA]

No I didn't make that implication at all, you just misunderstood what I was saying. FYI, most open source is done in low level languages, which are a pain in the ass. Sometimes its more of a tradeoff than its worth and closed source is merely more efficient in many projects.

no bias here is there!

[aletoledo]

I knew when I was looking into these comments that it would be a hotbed of Linux support.

I understand and recognize the point of the statement. There are lots and lots of flavors of open-source available and I would hazard a guess that a couple posters have created some of their own already. If your response is anything like "well I can reaarange the operating system code in open source" then of course you're biased, because the majority of people can't recode an OS.

Microsoft has a fairly straight-forward model: major release + patches. I suppose thats the same thing with OSS, but when OSS has a 5.3.2.7.4.8 release, I'll be damned if they didn't tag on the last couple of numbers just for kicks sometimes.

So I'll be the one to say it then, yes having 50 different versions of an OS (i.e. Linux) with a myriad of user created/customized apps is a lot more confusing than to just buy the most recent box on the shelf and then patch it to the latest level by clicking on a button.

An earlier poster described it best, that we should think in terms of Microsoft being a single distro. if all I every looked at was a single distro of Linux (say Suse) then things would be a lot easier from an OS standpoint to follow.

Re:no bias here is there!

[DarkDragonVKQ]

I agree, that's the main weakness of Linux. It has no one solid support and because of that there's no real incentive for a company to develop their product on Linux. If they had just one Linux OS that everyone worked on together (yes it would suck for freedom and creativity, etc..) then users would be more consilidated.

Re:no bias here is there!

[aletoledo]

having one Linux OS would truely be the answer, but then 10 years down the road I bet that people would start to lump Linux and Microsoft into the the same club and then start a brand new OS altogether!

I guess we'll always have to have an OS that highly diverse, yet is confusing to the majority of people and therefore a company such as Microsoft (or AOL if talking about internet) with always step in to make things a little bit easier for the average Joe.

Re:no bias here is there!

[DarkDragonVKQ]

Yep, because people always want to be different or stick out in the group. Which is fine in some cases but it causes hell for an OS.

Evolve or die.

[Anonymous Coward]

Evolve or die.

Ease of Use the Last Step in Open Source

[borroff]

I am an OSS consumer, not a contributor, but it seems to me that the first priority is getting the functionality right, and the ease of use features are tacked on afterwards. I know that's true for me when I create internal apps.

Right now there are tons of businesses thriving on just this fact. They use the open source tools, integrate them, and use a propietary binding layer to tie them together and make them easy to configure and use. Security appliances are the best example: Get an easy to use configuration of Postfix with some AV engines and Spamassassin, and you've got yourself a $5K box for the $500 hardware and your skull sweat, which can be endlessly replicated. Load balancers are another example.

Once these markets are established, then the OSS community moves in again, and commoditizes the new frameworks. Perl and Python fit very well in this niche, I might add.

Complexity? Maybe for an MCSE it's complex...

[Todd Knarr]

Well, some parts really are more complex. That's usually because there's more things to tweak. Setting up Samba, for example, is a bit harder than enabling file sharing on Windows, but then you can excercise more control over how Samba does things while Windows gives you the choice of doing it it's way or not doing it. More control = more controls.

On the other hand, a common lament about the "complexity" of Linux from the Windows side is "But on RedHat startup scripts go in /etc/rc.d/init.d but on Debian they're in /etc/init.d! How am I going to figure out which it is?!" (to which I usually respond "Well, if you can't check which one exists and just use that, you really need to go back to Shell Scripting 101.").

Re:Complexity? Maybe for an MCSE it's complex...

[dvNull]

From what I understand almost all distros now have their startup commands in /etc/init.d. /etc/init.d usually ends up being a symlink to whatever the actual rc.d directory is.

According to LSB: http://refspecs.freestandards.org/LSB_3.1.0/LSB-Co re-generic/LSB-Core-generic/etc.html

The only thing which is different is the actual service name. Take Apache for example, on redhat/fedora its /etc/init.d/httpd, on Debian its /etc/init.d/apache{1,2} etc.

Document management

[Flying pig]

Oh yes, I evaluate document management systems from time to time and, like you, I find myself wondering why people just cannot be bothered to organise the storage. Sadly, Laptop Road Warrior is the biggest obstacle - especially when he complains that his HDD takes forever to sync because of all the rubbish powerpoint he has created in 20 versions because he is so scared of losing his creative changes. But this is not primarily a Windows versus FOSS isue, this is a "how the hell did we let ourselves get into this mess" issue.

As for what these document management systems DO, mostly it is organise the "marketing collateral". Which could probably be better done using the old IBM buzzword sentence generator rather than Sharepoint.

Re:Document management

[Identifiable Coward]

I think you meant....

As for what these document management systems DO, mostly it is organise the "marketing collateral". Which could probably be better done by dragging the whole lot to the recycle bin!

Nope

[smcdow]

It's not that open source is too complex. The problem is that most IT and SW dev people are as dumb as rocks. And even smart IT people and SW devs are too chickenshit and/or lazy to take any initiative to self-teach themselves anything that's not right in front of them. It's all a part of the "Software Engineering Industry Going Down the Drain" zeitgeist.

Three different points of view

[Da Web Guru]

1) User -- For the average user, due to great advances in user-friendliness, learning how to use Linux is just about as difficult as learning how to use Windows. The names of the programs are different, however they all do similar things.

2) Admin -- This is harder in Linux. In windows, there are only a few differences between configuration locations between Win2K, XP, Win2K3, etc., and of course the registry is a common feature among all versions of Windows. In Linux, differing software package formats (rpm, deb, tgz, emerge, ports, etc.), differing configuration file locations and formats, differing startup scripts, differing authentication systems, and various other minor differences make managing one distro almost a completely different experience than managing another distro. Some distros have decided to use similar tools, however there is still enough variety to make things way more interesting than necessary.

3) Developer -- It is fairly difficult to release binary packages of your software when you don't know whether or not a particular library will be present on the target system. A package compiled for RedHat Enterprise may or may not work on a Fedora system, and will almost be guaranteed to not work on a Debian/Ubuntu system without a lot of library mangling. You can't just say "let the end-user compile the software on their system", because certain large-scale software products (i.e., ones that are very expensive and include commercial support) are not released with source code (for obvious reasons). Also, even if they did support multiple distros, they would have to deal with annoying miscellaneous issues such as whether a particular library file is present in /lib, /usr/lib, or /usr/local/lib, whether a particular distro changes library version numbers and doesn't contain a recent enough library, whether or not the distro has an autoupdate tool that can be used to update a single library without breaking anything else, etc. So, large-scale software manufacturers choose to restrict what distros they provide support for, due to the fact that they do not want to be responsible for supporting some random distro that they have not heard of or done any testing on.

Re:Three different points of view

[robertjw]

2) Admin -- This is harder in Linux. In windows, there are only a few differences between configuration locations between Win2K, XP, Win2K3, etc., and of course the registry is a common feature among all versions of Windows.

Nonsense. In Linux you have a standard, robust set of tools to use. There are, to my knowledge, only two commonly used startup script configurations in Linux distros. Windows has many differing software installers from homegrown stuff to MSI (which changes from time to time itself). The real advantage in Linux are the networking tools. It's much easier to help a user with her machine if you can telnet/ssh directly in and see what's going on - an ability is not native to Windows.

Developer -- It is fairly difficult to release binary packages of your software when you don't know whether or not a particular library will be present on the target system.

Actually, it's not. Offer a static build like Opera does. Not feasible in all circumstances, but definitely doable for many projects. Besides, DLL hell is one of the most renown problems in Windows. I've worked on Windows software releases. Attempting to support one Windows platform with all of it's potential library issues is hard enough. Throw in support for versions back to Windows 98 and (shudder) Windows Me, that becomes complex, especially if your application has networking capability.

So, large-scale software manufacturers choose to restrict what distros they provide support for, due to the fact that they do not want to be responsible for supporting some random distro that they have not heard of or done any testing on.

Not exactly. Large-scale software manufacturers (or any software manufacturer) is going to only test on platforms that will make them money. Testing is a lengthy, expensive process. Support even more so. If a vendor can control their testing and support costs by restricting them to the most lucrative platforms they will do so. It has nothing to do with what they've heard of, it has everything to do with what their customer base requires. If enough users start using the new Linux distro of the hour a successful vendor will need to support it in some fashion.

Re:Three different points of view

[trupoet]

And thats why they invented Gentoo

Ah, but that is the point.

[SmallFurryCreature]

You are concerned about MS customers being confused. MS is concerned about people who are not their customer being confused. They don't give a shit about the former. They are already hooked. Same reason roughly you don't court your wive. She lost that privilege when she said yes, she is now yours and you expand your energy on the cute girl in the office.

Isn't this bad news for brilliant developers?

[schngrg]

MS - time to face it, almost noone apart from you is making alot of money selling proprietary sofware (alone). The real cash is in services, services, services.

Isn't this bad news for the brilliant developers who can write excellent products which can solve real problems and that are so good (or small) that they don't need paid support?

btw, I think there are still enough software vendors making enough money with proprietary software too. Open source and proprietary software can co-exist, they don't have to fight each other out.

BS

[dskoll]

We sell commercial software. We have customers running our software on many versions of Linux (Red Hat 8, 9, Fedora Core 1 through 5, Red Hat Enterprise Linux 3 and 4, Debian, Gentoo, Slackware), Solaris and FreeBSD.

If you make yourself aware of portability issues at the outset, it's not a big deal to write portable software. Note that we only build binary packages for a subset of our supported platforms; source packages take care of the rest. (Yes, we distribute source, though not open-source.)

I don't think our support burden is any higher than it would be if we supported Windoze.

And to that end...

[maillemaker]

>The real cash is in services, services, services.

And to that end, it is most beneficial to insure that the software is as complex as possible, so as to make anyone who wants to use it dependant on those services, services, services. :)

Steve

Translation:

[Black Parrot]

"Our stuff is better than the competion because mix(fear,uncertainty,doubt,buzzwords,technobabble, lies,brainfarts)."

News at eleven...

Look ...

[Anonymous Coward]

I'll explain this very simply so the guys at Microsoft can understand.

The reason no-one develops proprietory software for Linux is that Linux users don't want to buy proprietory software.

The Linux OS is the pinnacle creation of a subculture that values a certain model of software "freedom" above all else. Proprietory software comes from a model and culture which is diametrically opposed to that. (Not from the vendor's POV, but from the Linux community's POV - only the Linux community has this "if you're not with us, you're against us" attitude, because only the Linux community has Stallman, but I digress.)

To say it's "hard" to develop a major end-user app for Linux is true, but not really any truer than for Windows, where you must support two fundamentally different operating systems - DOS-based 98 and ME, and NT-based 2000 and XP ... oh and Server 2003 too. And now Vista is coming down the pipe. All with different levels of OS functionality which you must emulate if it didn't exist in your baselevel platform.

I really don't see that developing for various versions of the Linux platform is that much more difficult in principle - but it is certainly different in degree, since there are an almost infinite number of flavours of Linux (everyone can have their own - that's the bloody point - so no guarantee can ever be made that a given shrinkwrap package will work on your system).

That's why it's stupid to "choose" Linux or say it's "ready for the desktop" and other such nonsense. Linux is not designed for the desktop and will never be ready for it. If you need to run proprietory software you need to buy a copy of Windows. But you can get a very long way without it, as millions of geeks have proven.

I really don't see what's complicated about this ;)

internal as well as external complexity

[Junks Jerzey]

The big win of open source is that many people can work on the same code, that work gets done faster as a result, that the code is more reliable overall. But for this to work outside of the usual suspects (Linux, Perl, Python, Apache), the architecture needs to be clean and dead simple.

All the time, working on proprietary code for a living, I see people making bad assumptions about how code works--even though all the original developers are right there in nearby offices--and there are often big whiteboard explanation sessions about little nuances of various libraries and low-level choices. A large percentage of the time these sessions end with the person thinking "Wow, I'm glad I didn't jump in and make that change without asking first."

This is exactly the same with almost any open source project of more than trivial size. The only way to combat it is to keep things as clean and blindingly simple as you can. Write as much in Python or Ruby as possible. Don't over-architect things (that is, don't make things pointlessly abstract). Ruthlessly trim features. Stay away from projects started by people with no engineering experience.

Just woke up and read open office

[TouchOfRed]

o/t here, but tired and groggily haivng my morning coffee I read "Is Open Office too Complex?". Quite.

Re:Open source is easier, if you know how to work

[archen]

If administrators have to resort to the reading the source when something fails, there is a problem.

I've often found this to be sort of a plus in a roundabout way. It seems that OS applications more commonly spit out specific errors about what is wrong into the logs. This often doesn't mean anything to you, but a search often finds someone who did have the same problem and poked through the code to figure it out. When you're an administrator of a criticle system you need it fixed. If MS just gives you "an error occured" message then when push comes to shove, you may very well wish you could just look at the code.

Although I'm pretty far removed from C/C++ now days, I've done search on error messages and come up with the actual code where the program generates it. Some times it's easy enough to tell what conditions are causing the problems without being a programming guru.

screwy logic and fast FUD

[rs232]

'It's challenging for partners to build competencies to support Linux, because you never quite know what you're going to be supporting,'

Like how? Since when does a customer demand SuSe or Redhat. It's a decision for the distributor what base OS to use. The customer doesn't even know or care what that is.

'You don't get that repeatable [development] process to build your business over time.'

Same flawed logic here. The developer sells the same basic system (with minor modifications) to a number of customers. The same as Windows is sold over and over again.

"One of the beauties of the open-source model is that you get a lot of flexibility and componentization. The big downside is complexity,"

Flexibility does not equate to complexity. Is Windows simpler to develop on because it is closed.

"We had to learn .. different versions .. to meet the demands of customers." Lim Han Sheng, IBS Synergy

Has anyone reading this ever had a customer ring up and specifically demand a particular Linux OS. I'm talking about a new installation and not upgrading an existing one. When I see someone quoted on the fastFUD web site I tend to suspect their impartiality. Lets see some more quotes from IBS:

"because Linux was keeping the company from charging for services rendered or charging less"

Has Malaysia gone communist and told no one about it?

"When a problem needed to be solved, Lim explains it was usually a case of having to ask friends or associates for advice"

Why didn't you ring up the Red Hat support desk?

"the tools available for MySQL were made by different companies, poorly integrated, and offered limited functionality"

"the difficulties IBSS encountered with Linux as a server .. relating to the display of fonts and font sizes, and the instability of the X .. IBSS could not integrate stored procedures into MySQL as it could with SQL Server 2000"

"IBSS is exploring the integrated innovation inherent in the Microsoft operating system"

ha haa haaa ...

People who think...

[Anonymous Coward]

...that OSS / Linux / [insert most hated software here] should be playing with their balls instead of computers.

Computers are complex so is advanced software, don't think you can make it simple... you can make it look simple but until we've got Star Trek like computers it will just be a illusion.

--

You may troll me... but it will not change the fact that this is the truth!

Re:People who think...

[Anonymous Coward]

Even the visest shit is just shit when it is unverified... I'll try again :
People who think that OSS / Linux / [insert most hated software here] is especially complex should be playing with their balls instead of computers.
Computers are complex so is advanced software, don't think you can make it simple... you can make it look simple but until we've got Star Trek like computers it will just be a illusion.
--
You may troll me... but it will not change the fact that this is the truth!

I call bullsit

[SmallFurryCreature]

You are saying that someone who knows solaris AND FreeBSD (or for that matter anyone who knows TWO operating systems (and no not windows 98 and XP)) has trouble when being shown a third that is based on the same ideas?

Yes there are various different ways to handle the startup of the various unix like systems out there but anyone who has already learned about TWO different ones will have learned a very important lesson. Things differ, so what.

I sometimes think it is a modern ailment because of windows click and play attitude. In the old days even if you used dos you had to learn stuff. You could just slide in a cd and play, you had to acess the A: drive then dir and find the install/startup file. You might have to edit your bootup files. In short you had to learn a fair bit about the system before you had any hope of using it.

Not handy and easy but when you landed in shit creek you at least had a paddle.

Windows has increasingly made it easy to just use the OS (or rather the desktop) without every learning about just what makes it all tick. Very good for endusers but when windows drops them in shit creek they instantly go under and often make the mistake of breathing in in suprise.

For us older folks we had to learn the hardway. The trick to learning an OS is realizing/accepting that you have to learn it.

If you drive a new car do you complain the gears work differently, are in a odd place or just don't respond like the car you are used too or do you just note the differences and adapt?

Your statement could work for a person who knows just one OS, wichever that may be, but never for a person who knows two.

OS'es are like learning languages. Once you got over the hardship off learning your fist foreign language and learned to accept that a different language is just different, just because they want to humiliate foreign devils (yes I learned japanese how did you know?), a third and fourth etc etc will be increasinly easy to learn.

How do I know? Well I am learning my fifth language and recently took up opensolaris after a dose of DOS/AIX/OS2/Windows/Linux(more distros then is healthy) and found it to be a breeze. RTFM, accept that things do not have to be the exact same way you expect them to be and adapt.

It is only a mess if you are unable to learn, to just accept that X does things in Y way. RTFM and a large brain should do the trick. In fact for me I think it is easier if things differ greatly then if they are almost 100% the same but not exactly. Like I keep using LS in DOS and DIR in unix. Never when I specify parameters, just the bare command. Exact same task, same looks, different commands. Different boot scripts. Easy peasy.

No You Can't Have Cake Too!!

[Anonymous Coward]

It does make it more difficult for a large company to develop for a Linux crowd in general

ANY company 'JUST NOW' getting into the development support for linux, is a little late to dinner, don't you think?

Yes the underlying platform remains static, kernel, driver-base, API stack... however, Linux is now more spliced than ever with regard to distro's. Any company joining the game now would HAVE to push support to multiple distros for thier app. If they don't, they risk being labelled only halfass behind OSS.

It just seems that this issue is touching on the wrong points. OSS, en masse, is a developer driven machine. Yes, some company's play a part in support and development, but the professional hobby coder is the driving force for this beast.

If OSS is to complex for some company's to consider, I have to wonder what value they represent when some industry leaders have thier business built on it, and entire government entities accept it with open arms and push it forward into thier environment.

I ask, who is selling who on what here??

Re:Open source is easier, if you know how to work

[Eivind Eklund]

Interesting, I've never had need to talk to a MS-SQL or any database developer to use any of the database products I've used. I've worked on some pretty complex databases and queries as well.

It's mostly a matter of convenience - the developers give better support, and the network of people around it give superb support. I suspect you have been trying to work with open source without knowing how to exploit the network around it.

As for increasing complexity (losing design) and lacking finish, that depends on the project, but yes, it is typical. In one way, it's a strength - open source tends to gravitate towards simple solutions that can be robust because those that try to push finish on it just don't have time. In others, it's a great weakness.

Eivind.

Is it a cardinal sin...

[Anonymous Coward]

Is it a cardinal sin, on slashdot, to not think "more choice and more configurability is always best"?

I've used the Macintosh operating system (both pre- and post- MacOS X), Windows (95 on), and a couple Linux variants (RedHat, Ubuntu). I've used a wide range of both consumer and professional software. I consider myself an informed computer user, but I'm not a what I would call a computing professional -- I'm a physicist.

My experience is open source software _is_ more complex from a user's standpoint than commercial software(complex in the sense of difficult to understand and properly use). Though I must say that software developed in-house at universities and national research labs can be even more unpleasant to work with.

But to cut to the point -- additional configurability and flexibility is only an asset if you regularly use or anticipate using those features. Otherwise, the consequences are steeper learning curves, more challenging troubleshooting, and often a less reliable product. I think a lot of open source software packages suffer in this way. (I say software packages in recognition of the 'many small programs' philosophy I see in linux.)

Now, the same complaint can be made of mega-programs like Office or Windows XP. But in these cases, there is at least some sort of market pressure to either get features (including UI) mostly right or not to do them at all. That market pressure doesn't exist in open source software. In fact, arguably, a business model based on providing services _supports_ software complexity.

Re:Open source is easier, if you know how to work

[Eivind Eklund]

I agree that startup cost is an extremely important point for sales, and that open source projects often suck at it. I believe it's also important inside open source - for instance, I believe a large part of mysql's success compared to PostgreSQL is startup cost.

As for "RTFM, idiot" - I hear people tell of this reaction all the time, yet I've never seen it. Maybe it's to do with what projects I choose to deal with - I use FreeBSD instead of Linux, and do work with PostgreSQL/Perl/Ruby (and, due to costs of migrating away from it, MySQL), and I've never seen that reaction in those communities. I've seen the reaction "man XYZ" or similar (pointing at the appropriate documentation) regularly, but that's generally of help to the person being told what to read. And if the person has a problem understanding the documentation, we try to help and update the documentation as appropriate.

Eivind.

Re:If Open Source software is so complex to use...

[QuietLagoon]

I was not aware that slashdot was an OS vendor.

Things which are poorly understood appear complex

[Colin Smith]

Once you understand how something works it isn't particularly complex. What he's saying basically is Linux... Doh!

Linux is Linux whichever distribution. There may be things in different places now and again or may be missing here or there but lets face it, not remotely difficult to fix.

Yes.

[jafac]

Open source products are much more complex to integrate.

But then, with closed-source COTS products, you often run into vendor-lock-in scenarios that bring even worse headaches. Some of those headaches result in Security pain, as you jump through hoops with forced upgrades and hacks to keep things working at the application/environment level without any hope of capability to fix problems (like LUA bugs on the Windows side).

I'll take the Open Source configuration complexity any day.

Registry

[guisar]

Not have a central configuration database is one of the reasons I much prefer Linux. A series of PLAIN TEXT config files in /etc is fine, per user config storage is excellent. A central database has to be protected from inadvertent changes, reducing my ability to configure things how I chose. Second, the more the components of an application are spread throughout the filesystem the more difficult system admin and recovery becomes. Why shouldn't I be able to copy an application from one machine to another, just the binary file and have it come up at least in some default state? I think this should be a goal for all applications and a central configuration database makes this just that much more difficult. I won't even talk about the cruft and corruption issues a central database brings with it.

Information Theoretic Limitations of Formal System

[Anonymous Coward]

If your operating system is capable of solving many problems (i.e. proving many theorems since after all, computer programs are proof systems), that is, if you system has a large deductive closure, then by Algorithmic Information Theory we know that it must be complex. Therefore, calling a system complex in a perjorative way is ignorant. Correct me if I'm wrong.

show me the money

[Anonymous Coward]

I am a network Engineer and I think Whiney made 1 good point in the first post - the money is in services - if I have to choose a technology to become proficient in supporting, or a OS to write apps for, I'm going with the market leader. It's about the money. I support MS because I have worked on their products since DOS and I know it. I support Cisco because that's what interests me the most and I have become highly certified in networking/security/voip. Complex - yes, worth the money - yes.(again, a market leader) I use open source for the little things here and there and I agree that in most cases, open source is more confusing and takes longer to setup than the equivilant market leaders option.

My god, it's full of stars...

[Wolfger]

So Open Source is bad because it gives customers too many choices? That's a hard sell. I mean, if you or your company buy Linux from many different distros, you will have a headache on your hands, sure. But most businesses (and people, for that matter) will use a particular distro across the board, and then the situation is no worse than if you chose Microsoft as your platform. Single vendor, single platform, consistency. Multiple vendors, multiple platforms, inconsitency. Open Source or not has nothing to do with it.

Re:My god, it's full of stars...

[zpok]

For enterprise market, I'm with you, for consumers, I'm totally not with you. I loath the complexity of linux. Not linux, or foss. I'm in love with the idea, but leave the using to those who like it.

btw, to all readers: If your reaction is "my god, you're simple", just assume you're dead-on and don't wear your keyboard out ;-)

Re:My god, it's full of stars...

[Wolfger]

Well, I can appreciate where you are coming from, but I can't really say I agree with you 100%. There are complex Linux distros to be sure. Then there's Linspire, SimplyMEPIS, and probably some others that are every bit as "plug n play" as Windows. Or maybe that's just me....

Re:My god, it's full of stars...

[zpok]

Two that I haven't tried yet. I'll try them out in a month or two when I have the time.
Not that I'm likely to switch, just am a consumate OS junkie, want to at least try them all.

For shrinkwrap he is right

[5937]

As enduser i can take a windows-program and run it everywhere. In Linux i need a platform-specific build.

Thinking i am a user: that means distro-supported programs are simple to install, but other stuff is prohibited. "make install", some path wrong, fails. I have no idea where it wrote in all this system-locations. All the root-owned areas i know nothing about, and better dont touch.

Thinking i am a shrinkwrap-developer: that means, i better make packages for each distro, and get that in the official repositories. Which means a lot more expertise than just clicking "compile". At least it looks like that when starting. Some days learning for somebody. And even different defaultfonts can break an app. And the social issues of getting it in the official repo.

What to do about it?
- Use some vm, like java or wine, which already encapsulate the differences.
- Let "closed source" apps be supported by the distro-communities and let them do the compiling. Would need some kind of job-market, where a developer could quickly find assistance. Still effort, but fixing a makefile is much faster for a distro-insider.

Unix support far more confusing than Linux

[freeweed]

Saying you support 'Linux' is silly. It's like saying you support 'UNIX.'

Can't agree enough.

I had a consulting gig where I had to support Solaris, HPUX, AIX, and an ancient SCO Unix box. Those 4 Unixes were FAR more different than say, Red Hat, Debian, Suse, and Gentoo. Even the basic command structure often differed.

Different locations for config files? Different desktop environments? Child's play.

Try switching between machines where the most basic things are entirely different. Ifconfig is a notable favourite. The command switches, and the output, are more different between commercial Unixes than they are between say Linux's ifconfig and Windows' ipconfig. Hell, IBM has an entire set of system management tools all wrapped up in good ol' smitty - which no other Unix has.

I won't even get into library support, boot process, or development environments. Knowing the 5 or 6 major Linux distros, and their differences, is trivial compared to a mixed-breed Unix shop.

One Word

[cab15625]

Regedit

Seriously, has this guy ever tried to fix anything involving the use of regedit. Or tried to uninstall anything more complicated than minesweaper? Complexity is no basis for avoiding OSS in favour of MS.

Re:One Word

[Dark_MadMax666]

Have you? What is so difficult about it? *nix lack standards ,ease of use for user and centralized administration for admins. Saving couple hundred bucks on licenses is not worth the hassle in 90% of cases - even with all MS problems (which are proprietary interfaces ) . If you manage corporate environment with dozens of applications ,hundreds servers and network devices and appliances of top of numerous desktop clients you do want unified environment which for most of the time " just works" TM .And guess what? -it ain't OSS stack. Sure tomcat/Apache here and there, some*nixes on a few appliances ,maybe *nix based firewall , few nifty OSS utilities but backbone is AD , Exchange, MS SQL and W2k3 servers.

Re:One Word

[cab15625]

Yes, I have. Mostly on our test boxes. Test some software for some user. Find it makes the system unstable. Uninstall it (several reboots later...) Test the system and find it didn't get it's stability back. Call tech support. Oh, well you have to go to regedit and remove key H_BLAH_BLAHBLAH/FOO_195/12348907610785/bork/pop/ou ch451/poop. Why didn't the uninstall do that? Why isn't it called something more indicative of what it does?

As for "just works" well, I've had some fun with beta software on linux that didn't "just work" but apart from that, I can't think of an official release that I've played with that didn't "just work". We have a pretty diverse (computer wise) setup out where I work. We use OpenBSD for file,print,mail,web server. We connect to the server with w2k, XP, OSX and several flavours of linux. It's reliable, stable, and (OMG, ) documented! Shelling out the extra cash for almost non-existent and all too often incompetant tech support and a couple CD's of bloat would not have been worth the effort for us. The whole "just works" thing, in my experience, only applies as long as you only want to do things exactly the way MS dictates that you should want it to. Everything is simple as long as you don't need to tweak it for a specific need.

I stand by my main assertions. Complexity/simplicity is not a good reason to switch. Microsoft has no business pointing at other systems and whining about them being complex.

You may understand the layout of a MS system pretty well, and therefor, don't find it to be as complex as people who don't have to deal with them on an daily basis. But they're no less complex than OSS. Just different. If you had a clue about how to maintain *nix/OSS, you'd find the assertion that OSS is too complex as laughable as you found my original post annoying. I am not a Windows expert. After the few times that I've had to deal with the guts of a Windows computer, I really don't think I want to be. They're complicated, finicky, and just plain obscure once you have to do anything at all interesting.

Re:One Word

[Dark_MadMax666]

Well in this example the problem lies in buggy software ,not with OS per se. I had my share of troubles with some bad software (lately symantec server product - veritas and corporate AV tend to have those problems) - but it purely ISV lacking common sense and quality control. And it was solved pretty easily - I wrote a vbs script to automate removal of several dozen entries and run this script from my workstation to execute and and clean up all affected machines

  In absolute numbers there is more of bad 3d party apps under windows -mostly because there is just more software for MS platform.

What really makes my day is that I can write and admin script using well documented API and it will "just work" for windows platform - most of my script work for all win machines in the office (from win2k pro sp4 to w2k3 x64 r2 enterprise) without any modifcation or exception handling code. -I couldn't dream about something like that under linux , as none of the distros have stable management infrastructure ,except maybe redhat (and it is lacking comparing to windows) . -Linux distros cant even agree on same place to keep config files ,let alone provide consistent interface/API for them (and windows not only does that - it provides it in centralized manner).

  MS really does an excellent job providing unified management and administration interface to its whole stack (not only OSes but AD, Exchange ,SQL -you name it ) and maintains this interface well . As an example - WMI provided new ways to manage computer trough scripting . It is really excellent tool - but win2k didn't support it initially . yet M$ added support to it so I dont have to think that half of my machines support WMI and half don't.Another example is your registry -I have command line way , script way ,api way to manipulate registry .Including batches of remote machines - and those machines can be diffrenet "builds" (w2k ,winxp,w2k3). For linux distro i would have to tailor the script for each specific distro , ohh ,and btw worry about setting up infrastructre for secure remote authenticaion as well ( all modern winos do that out of the box).

  Course it is not all roses there , but compared to OSS I do prefer consistency and if you want "simplicity" to mishmash of OSS any day.

p.s. And I do have a little clue about *nix/OSS- I do maintain scores of apps using OSS stacks (firewall/spamfilters,tomcat/jboss app servers) . -I did play around debian/redhat etc and we have several workstation running suse. In small niche OSS may work very well - but as infrastructure...- It lacks vision and design for that, You cant have such thing when linux cant even maintain stable API/binary compatibility .

Re:One Word

[cab15625]

Well in this example the problem lies in buggy software ,not with OS per se. I had my share of troubles with some bad software...

Fair enough. But the counter to that is that it should not be so easy for flaky 3rd party software to mess up the OS. It should also not be so difficult to track down turds left by bad software. This is where my original complaint with regedit came from.

What really makes my day is that I can write and admin script using well documented API and it will "just work" for windows platform - most of my script work for all win machines in the office (from win2k pro sp4 to w2k3 x64 r2 enterprise) without any modifcation or exception handling code...

Sweet. I wish I had that skill with windows. But I do this with Linux all the time. Scriptability is one of the main things that makes *nix (including Linux) so nice to maintain. Maybe it's just a difference in style or a lack of VB skill on my part, but I've found scripting to be problematic at best underwindows. I do have to agree with you on the config files thing. It can be frustrating with some apps/distros to try and find where the config file is. However, most of the big ones come with both a gui and documentation if you don't know how to use "find" and "grep". In the end, it's really no worse than trying to hunt down some obscure key in the registry. If you want to stop a program from autolaunching when a user logs in, under windows, how many places do you have to look for it in the registry? Last time I had to do this there were almost 10 places to look (think about symantic being cute and placing stuff on the task bar, start menu, and scattered around registry.) I don't see any significant difference beyond one being what you are used to.

MS really does an excellent job providing unified management and administration interface to its whole stack (not only OSes but AD, Exchange ,SQL -you name it ) and maintains this interface well ...

Most distros are also pretty consistent internally. Even between distros there are a lot of standard libraries. This is why, eg. Firefox, Samba, LAMP, SSH, and a whack of other software are able to work on my favourite little niche distro (Slackware) just as well as they do in the major ones like RH, Debian, Ubuntu, Suse, ... And if something doesn't work right away, I know where to find some documentation, even if it has to be source code, to find what the problem is, and how to fix it. I've very rarely gotten good help from MS, either in the form of documentation, online help, or over the phone. This is probably a large part of my frustration with using various generations of their OS.

p.s. And I do have a little clue about *nix/OSS ...

Sorry, it's been a while since I layed into someone like that and you didn't deserve it. As far as lack of vision goes though, in my experience, it is MS and the various incarnations of Windows that have demonstrated poor design and lack of vision ... and been overly complex and frustrating to work with as a result. If their system correlates well with the way your mind works and you can find your way around easily enough, then keep using Windows. To me, it looks like an insane mess compared to the OSS and *nix stuff that I prefer to work with. Both systems have their complexity, but OSS, in my experience, has done a better job of providing a way for the user to get through it.

Microsoft spreads FUD. Film at 11.

[mr_mischief]

Wow. Software development for a particular group of OS distributions is complex. What news.

Come on, if anyone's a software developer instead of a fanboy who does copy-and-paste until something compiles then they know how to manage complexity. Software development is the management of complexity, after all. You take a big problem and break it into tiny steps a computer can perform one after another. In many cases you even figure out what things in that set of steps don't depend upon each other and break the steps into multiple threads or multiple processes, sometimes on multiple machines.

The APIs in common for Suse, RedHat, Gentoo and Ubuntu are in my experience much easier to find and stick to than those that Windows 98, Windows 2000, and Windows XP have in common. But then again, I do most of my software development on Linux. When I do it on Windows, I tend to use libraries to make it look more like Linux so I don't have to write to Windows APIs directly. I would suppose that if I was a Windows software developer, I'd find Windows more familiar and try to make Linux feel more like Windows. I'd probably write to Wine if I was coming from Windows and wanted stuff to run both places, just like I write to tk, SDL, or OpenGL now to get stuff to run on both. Where's the news?

The short answer...

[burnin1965]

No.

I suspect that in most cases the real issue is a lack of understanding of the fundamentals underlying the technology people are trying to use. While a nice UI and preconfiguration may make it possible for someone with limited knowlege to be capable of installing and configuring a complex software package it is a poor solution to the lack of knowlege.

Having said that I will also go so far as to say that open source has a better solution to the complexity issue than Microsoft's wizardized software packages, LiveCDs. You can download a live linux CD for just about every solution you need, Desktop, Router, Webserver, Media Player, etc., etc. No need to go through Microsoft's complex software package installation and registration process, just pop in the CD for your application and boot, done. ;)

burnin

X11 chattiness

[steve_l]

Yes, X11 is a chatty little bunny. it was written for the lan, not the long-haul link, and one of its design goals was to support diskless client workstations, at a time when HDDs were expensive. Nor is it good for gaming, which is why linux 3D gaming bypasses X.

But funnily enough, one little benefit of X11 is that a Unix GUI application can run on a different back end from the front. Which lets you run a legacy solaris app on a legacy solaris server, with your ubunto laptop getting the front end, with much better integration with the desktop than you get with VNC.

freedom is scarry, but

[anwyn]

Free software creates choices, some are afraid they might make the wrong choice.

  "If ye love wealth better than liberty, the tranquility of servitude than the animated contest of freedom, go from us in peace. We ask not your counsels or arms. Crouch down and lick the hands which feed you. May your chains sit lightly upon you, and may posterity forget that you were our countrymen!" Samuel Adams

Config file formats vs. Registry

[mrchaotica]

You're right and you're wrong. Having every program use its own format is stupid, but so is creating a Registry.

The good solution is the one Mac OS X uses: have text config files for everything, but put them all in a standard format (plist).

By the way, your assertion that UNIX apps all put stuff in different directories is wrong -- well-behaved UNIX apps should follow the Filesystem Hierarchy Standard.

Re:Config file formats vs. Registry

[einhverfr]

I have witten a set of utilities to store text config files in an LDAP database.

Cool. Now, I have a network transparent registry with different formats for each program. SOunds like a pain, but it is still easier to manage than the Windows registry...

More complex for who?

[sketchman]

I use Open Source everything on my computers, and I migrated from Windows98. I have to say, even when my Linux apps do have a problem, the need for the fix is almost always not the OS's or the program's fault, and it's easy to fix. Usually it's just a file download, and it was my fault that file wasn't there in the first place, because the OS told me that I needed the files when I installed the program. And if there is a problem in code and I don't want to wait for the dev team to fix it, I can try to fix it myself. Not an option with closed source, at least not legally.

So, for me at least, Open Source is simpler and works better than closed source programs.

I can't imagine that process and situation being complex for anyone.

Compared to what?

[Alpha830RulZ]

The gentleman said, "Gavin noted that the flexibility of open-source software in meeting specific business needs also means systems integrators and ISVs have to grapple with complexity costs. 'It's challenging for partners to build competencies to support Linux, because you never quite know what you're going to be supporting,' he added. 'Customers who run Linux could be operating in Red Hat, [Novell's] Suse, or even customized Debian environments,' he explained. 'You don't get that repeatable [development] process to build your business over time." Over the last 10 years, if you developed for the MSFT environment, you've your code obsoleted three times, you've had to change out your client platform four times, and you've had to redevelop your server competency three times. You have to test for at least three client environments, and you have to justify to your clients and users a significant investment in OS upgrades. How is this different? As a tech middle management type who has tried to remain technically current, I have found it easier to get current and stay current in the Linux world than Windows. Sorry, Gavin, that dog won't hunt.

Even more so for Vista

[phorm]

Exactly how many flavours was Vista coming out with? Six? Seven??

MS: Linux sucks. film at 11

[wardk]

say it ain't so!! Microsoft has unkind words about linux?

stop the damn presses, we have a font page scoop!

As opposed to Windows and OSX?

[clambake]

Where you could be working with the EXACT SAME SOFTWARE and yet find the "rules" to configure it change because you have the "professional" vs the "home" version, or because you have 10.3.9 instead of 10.4 (0.0.1 away in version, 10,000 light years away in reality.)

Choice is good.

[twitter]

This is one of the fundamental things that Linux advocates rarely get.

We can take it that you are not one? That's too bad.

From a user perspective, there are about as many differences between Red Hat Enterprise Linux and SuSE as there are between FreeBSD and OpenBSD, or Solaris and some other SysV variant.

That sounds right, but it's dead wrong. Most people like having a choice up front and then many choices after install. The differences between the free and non free worlds are, as usual, huge and mostly in favor of the free world. People in the free software world get to try out and chose the interface they like. Once they have chosen they are only forced to upgrade by compelling new features and the overall experience is much more consistent.

Free software, once a distro choice has been made, is much more consistent than Windoze from a user stand point, and I'll even say that there's as much or more consistency between distros than elsewhere. The differences between RHE and SuSE are not all that awful because both pull from the same upstream sources. Konqueror on RHE does about the same thing as Konqueror under Debian and so on and so forth. The main difference is in administration styles, which the average corporate user could care less about. It's much less annoying than the pointless differences between Win2k and XP or XP and Vista, ad nausea. One "version" of Suse is much like the one before it, which is something the user appreciates much more than having transparency or some other useless toy. I don't even want to talk about the differences between non free GUIs like Apple, Solaris and Windows. Each has it's own version of window manager and users of one are clueless on the others. With free software you can at least have the same window manager and the window managers can share tricks. Non free GUI makers get smashed for "photocopy" when they try that.

Saying you support 'Linux' is silly. It's like saying you support 'UNIX.' Saying you support RHEL makes sense.

Well, that makes sense. Everyone has their specialty, except that poor guy at the help desk who's supposed to know every version of every free and non free thing ever written. I'm not going to tell people I can help them out with their Windows box. I'm most comfortable with Debian desktops and a few of their derivatives. Between Google and that, I can operate and stumble through other distros.

Re:Choice is good.

[nasch]

We can take it that you are not [a Linux advocate]? That's too bad.

Why is that too bad? Must everyone be a Linux advocate?

Most people like having a choice up front and then many choices after install.

I would say YOU like having lots of choices. The average PC user? I'm highly skeptical. I don't think they want to choose between Windows, Apple, or umpteen Linux distros. As it happens, they're not even presented with this choice, so it's just hypothetical. But I doubt your average guy off the street would be happier to walk into Best Buy and see the same computer for sale with six different operating systems on it, and be asked which one he would like. He wants to just buy the computer and take it home. Same thing with software - he doesn't want to choose between different browsers, different office suites, etc etc. He wants to just run THE program that does the task at hand.

Not even proprietary is hard.

[twitter]

Even proprietary software can be packaged coherently and ignore the rest of the distribution. Witness Star Office, which came as a big tarball that installed as easily on Debian as it did on RH, Solaris or Windoze. There are lots of other examples that prove it can be done easier than you can develop for the DLL and version hell of Windoze. Stuff on *nix works.

Re:Not even proprietary is hard.

[llefler]

There are lots of other examples that prove it can be done easier than you can develop for the DLL and version hell of Windoze. Stuff on *nix works.

DLL hell is just as prominant on Linux as it is in Windows. They just aren't called DLLs. The only thing close to a silver bullet is static linking, which is what some commercial Linux vendors have done. Works on Windows too.

Re:Not even proprietary is hard.

[iced_773]

Let me tell you a little story about DLL Hell and Linux. I was installing Kubuntu 6.06 on a system with a DWL-520E wifi card. (This particular model of NIC won't work with Linux out of the box, since it has no Flash ROM and thus the firmware needs to be loaded into RAM from a .hex file on startup.) So quite naturally, when the LiveCD booted, it didn't recognize the card. Therefore, when I installed the OS, it didn't find the need to bring its autoupdater with it. Of course, I got the card working, but trying to apt-get update firefox was a real chore. (I would use Konqueror but its accessibility options are limited and I have a visual disability.) It needed all these updated versions of packages and didn't seem to want to get them itself. I also tried downloading the Debian package, but I had the same problem installing. Suffice it to say, I didn't keep that installation, but its dosfsck at startup cleared up a small problem I was having with RSS on my Windows partition...

My point is that while "DLL Hell" is real, but Linux has its own Dependency Hell. Before you judge someone/something, make sure your own hands are clean.

Some are missing the point

[baggins2001]

One of the big issues is versions of platform. He said it right up front RedHat, Suse, Ubuntu. Which one are we developing for? If that isn't bad enough very few have a stable version which they will support for 5 years. Yes there are server versions for each one of these which are supported for 5 years plus, but most of the linux community is using the latest and greatest flavor of Fedora, Suse, Ubuntu, not the 5 year stable version.
Who wants to develope for a constantly moving target? Especially when the total market is less than 5% of users.
Now let's go back to RedHat, Suse. They have stable products, but look at the cost. The last time I looked a basic server version from RH or Suse, cost $349 per year over 5 years thats $1745. Of course both of these come with a lot of capability. But more than an equivalent cost of a MS server? Yes there are a lot of hidden costs in a MS server, but you have to really struggle most of the time to get the bean and pointy heads, beyond that initial cost. The concept of CAL's usually totally befuddles them.
So for the most part you have a small percentage of the market share using linux with a version that is going to be stable for 5 years. The remaining desktop users are changing their version of Linux every 6 months.
So who the hell in there right mind is going to spend a shit load of development cost on a product that runs on a platform that nobody is probably going to be using 1 year from now.
Currently I see it as linux is currently suffering from chicken and egg syndrom from about every direction. A large number of people won't use linux because they don't have the apps they need. No one will develope the apps they need on linux because not enough people are using it.
No companies develope drivers for devices, because not enough people are using it.
There are not enough people out there to support linux, because there aren't enough people out there using linux.

But, could you imagine what the newsgroups would look like if 15% of the desktop market were linux, how about 7%. 7% would probably double the traffic. 15% would probably triple it and I hate to think about the flame wars started between know it all's and newbie's. That's another hump we have to get over. Not everyone has thick enough skin to get their information from newgroups.
We still have quite a few hurdles to make it into the main stream. I just wonder how many of us are going to tough it out for the next 5 or 10 years for that to happen.

pot, meet kettle

[Tom]

Is Open Source too Complex?

Is this from the company that already admits it can't document its own products because they're too complicated? (see EU case)

Re:Open source too complex? Look at windows itself

[clm100]

Because on linux you never end up in rpm dependency hell. You never have to compile a kernel to use, for example, more than 4GB of RAM. You never have /dev identifiers moving around.

I agree that windows can be a pain, but the modern windows 2003 server environment is actually fairly easy to use, in part because of the low complexity. Microsoft will never have apt-get/emerge/package manager of choice and all the power that provides, and it will never let me compile a kernel for some custom purpose. But then I generally don't have a need to that on windows because they provide a flexible enough kernel built in that a few config changes make it adapt.

Remember guys, just because Windows isn't the One True Operating System doesn't mean that linux, in any of its forms, is.

Complexity, shmomplexity

[Tony]

Or with several different versions of its own operating systems, each with its own quirks and incompatibilities?

Complex software systems are, well, complex. Rolling out an Active Directory infrastructure is a pain in the ass. It's just part of the business.

Suse isn't that different from Red Hat which isn't that different from Ubuntu. The concepts are all the same, and the web is there for anything that might be different (layout of init scripts, location of default configuration scripts, /etc).

It's all just the same old FUD, wrapped up in new banana leaves and presented as haute cuisine. What else should we expect from an employee of Microsoft? Refreshing honesty? Yeah, I can see that: "Hey, everybody! Computers are fantastic! Yes, Microsoft has fucked up the industry and set us back a few years, but they are still great, and we're much, much better than we were. So, go out an use computers, and make them all work together to do the job you need. Free software is great, because you can customize it and it gives you the freedom you need, but hey, Microsoft products are great, too, because they are flashy and pretty, and the admin GUIs really work, and they don't crash like they used to.

"Oh, and Bill totally rulez. Jobs completely droolz."

Seems to me the complaint...

[DragonWriter]

...isn't so much that "open source software is complex" (e.g., any individual Linux system is more complex than any individual Windows system), but that, for instance, the range of things covered by the label "Linux" is far more diverse than the range of things covered by the label "Windows", so its harder for someone to develope the knowledge to say with credibility that they can support "Linux" than it is to develope the same knowledge with respect to "Windows".

Which is pretty clearly true; but its just as true, if not moreso, if you want to say you support "Unix" compared to "Windows"; it has nothing to do with "open source", per se, or complexity of software. It has to do with supporting more different things, and "Linux", without further specification, is more different things than "Windows".

No, X11 has lots of fire and forget commands

[2901]

Quoting from the Xlib Programming Manual by Adrian Nye

A protocol reply is sent from the server to Xlib in response to certain requests. Not all requests are answered by replies --- only the ones that request infromation. Requests that specify drawing,for example, do not generate replies. When Xlib receives a reply, it places the requested data into the arguments or returned value of the Xlib routine that generated the request. An Xlib routine that requires a reply is called a round-trip request. Round-trip requests have to be minimized in clients becasue they lower perfomance when there are network delays.

The designers of the X protocol were aware that round-tripping would make the protocol perform poorly in the presence of network delays and only put in the replies that are logically necessary.

I think there are two weak spots.

First, there is no provision for local echoing. You press a key, the event goes to the client, the client says:draw a character. Now the events and the requests are all queued, so several key presses can be in flight at the same time, but it is undesirable to have a round trip delay before a character appears in response to a key press.

Second, it is tempting to knock together a crude client that responds to every event by redrawing its entire window. This is easy to code, just update the clients internal state, clear the window, and call the GreatBigDisplayRoutine. Much easier than working out what changes need to be made to the window and just doing that. It works OK locally, but generates lots of network traffic. There is a case to be made that X11 is not at fault at all here. X servers are designed to support oodles of sub-windows, nested as deep as you want. If you take advantage of this much of the client code is simple, an event goes to a little sub-window that can quite reasonably be re-drawn in its entirety.

I've actually been studying (and writing about) CLX, which is the native Common Lisp binding to the X protocol. The native C binding to the X protocol is called Xlib.

Notice that X11 is an old protocol, from the bad old days when computers were small and slow, and Ethernet ran at 10 Megabits per second. It is reasonable efficient because it had to be.

It isn't about distros -- it's about mindset.

[Richard Steiner]

Actually, it should be relatively easy for a software vendor to provide statically linked binaries that install in /usr/local on just about any Linux system using a standard makefile or shell script (or pretty installation routine if they like).

That would remove most library dependencies as well as most of the filesystem idiosyncracies that one finds from distro to distro.

Monolithic apps require monolithic testing.

[Richard Steiner]

That's why Microsoft has such humungous testing labs -- because the software they have to test is designed as a large monolithic entity, not a highly-modular entity, so testing that software is a humungous undertaking.

If their software were written in a more modular manner, such large testing facilities would be far less important to them (and would perhaps be completely unnecessary).

It's the Cognitive Load

[kap1]

You can only hold 7+-2 things in your short term memory at once. This limits your ability to learn and problem solve. Complex environments increase the number of concepts you have to juggle in your brain at once -- the cognitive load -- and make errors go up and productivity go down.

This isn't a problem that is unique to open source, of course, but it isn't one that is focused on a great deal, with the exception of frameworks like Ruby on Rails.

*Maybe insightful. *Definitely off-topic.

[Mateo_LeFou]

If you read TFA, you see that Gavin's talking about developing open-source software applications as a business, not using said software. It's too complicated to sell software targeting Linux et. al. because there are so many version. I read his thesis statement as:

Since open source software is not a monoculture, it'll be difficult for an ISV to crush all competition and establish decades-long monopoly rents

This is a "drawback" that I, needless to say, can live with.

In short, the first post on this article pretty much could've been the last.

Linux Standard Base

[retro_alt]

Isn't this this similar to what the LSB distribution certification http://www.freestandards.org/docs/FSG_Imperative_W P_Public.pdf is trying to ensure? That If someone is using LSB-certified distributions, that it is guaranteed to have certain components that are needed for developers trying to reach a broad linux demographic? And as distributions adopt futher revisions of LSB, things can only get better? The biggest hurdle I see is the continuing battle over gui-toolkits. But even then, there are projects trying to make a way for developers to make gui's that will use the toolkit of the user's desktop environment, the Portland Project, I believe.

What Sort Of Product Do You Have?

[Anonymous Coward]

Can you name it? If not, is it a client-server app with rich client or a web app?

Re:What Sort Of Product Do You Have?

[swillden]

I don't think I can name it, actually, since I wrote it under contract and don't have the client's permission to disclose it. It wasn't client-server, or a web app. It was a set of applications that work with a smart card to provide strong authentication and secure PKI services. You use your smart card to log in, to unlock the screen and to authenticate to various services, local and remote. As a test of how hard cross-distro Linux apps are to build, I can't think of a better test. It touched hardware subsystems, authentication/login systems, and regular application-level tools as well. I should note that it actually wasn't only for Linux, I supported Solaris as well, and that was much more difficult.

Utter Total Compatibility With All OSs

[Anonymous Coward]

We have achieved utter total cross-platform compatibility in the following manner:

1. Write all apps in Microsoft QuickBasic for DOS environment,
   
2. Installation mods (mostly shortcuts and file permissions) with shell scripts to detect OS: Win95, Win98, WinNT, WinXP, Win2000,
   
3. Use WINE on Linux.
   

We think it's funny too!-)

Re:Open source is easier, if you know how to work

[mvdwege]

My experience with PostgreSQL some time back was much worse than any experience I had with SQL Server. The online documentation failed to match anything resembling the actually installation that existed.

That's weird. Every PostgreSQL point release I've seen, the PG team updated the documentation at release time. Whoever did the install for you fscked up, as the docu for the installed version should be available. If it's a source install, someone forgot to download the current docs, and if it's a distro package, someone forgot to install the documentation package.

Mart

Repository ease of use?

[tepples]

I use FC5, and I use third party yum repositories for any software not officially provided by the main Fedora repository.

But is it easy for a home user of Fedora Core 5 to add a software publisher's repository? And is it easy for a publisher to charge for access to that repository?

Pray what does Mandriva lack?

[rmm4pi8]

It seems to me painfully obvious that Mandriva 2006's installer is light-years better than Windows, and entirely graphical. I've heard good things about SLED10 as well. Sure these distros may not have the market dominance of RH/Fedora or the coolness of Ubuntu, but they've been around for a long time and polished things to a high sheen. Try them out.

Are users too simple?

[Hosiah]

Open Source software isn't nearly complex enough. We used to be assured that computers were functioning at full capacity. Then along came Microsoft cripple-ware, and all of a sudden computers were nothing but gaming consoles with WebTV functionality grafted on. We had to wait until a free Unix to run on our home machines before we could actually COMPUTE again. For the people to whom compiling your own kernel is just too much to handle, then COMPUTERS are too much for you to handle - go get yourself a dandy little cell phone to surf the web and a swift little XBox to play games on. With my blessings.

Computers are powerful. I mean HOLY-SHIT-I-DIDN'T-KNOW-THAT-WAS-POSSIBLE powerful. They haven't come near potential yet. They will evolve in leaps and bounds. Now there is the box that I need to program in six languages, produce 3D ray-traced scenes, optimize work-related tasks with an ever-growing library of scripts that approach closer to artificial intelligence with each new patch, catalog all useful human knowledge, and generally use office apps and accounting apps and earn my living on. And there is the box for Joe Sixpack to play games on, or for gramma to email her bridge friends on. At some point, as our needs evolve further and further apart, doesn't it seem just plain illogical to call them both a computer? One of us needs a new name.

But of course, Joe Sixpack is in the majority, so he will have the floor. And Leet Egghead will use his wits to circumvent the system in the background like always and keep using his "complex" software to do "complex" things. What a pointless question. Hey, is rocket science too complex? Is organic chemistry too complex? Is brain surgury too complex?

Simpler Licensing though

[Gumber]

Open source environments may or may not be more complex than a pure microsoft shop that upgrades in lockstep of Microsofts product schedule, but they will never, ever, reach the licensing complexity of Microsoft's products, which microsoft has intentionally obfuscated because they know that a confused customer will often pay more to be on the safe side, rather than risk a BSA audit.

The article is right

[GWBasic]

Please don't mod me as a troll, from my experience the article is 100% correct.

In 2000-2001 I worked for a company that made telecommunications equipment that's used in phone switches and the servers that tell you "Push 1 to..." Their equipment, (and bundled software,) had to support 99.999% uptime. (It translated to roughly 5 minutes of downtime, per server, per year.) The company had a history of supporting Windows NT, Solaris (both x86 and Sun machines,) and Unixware. (They also supported OS/2 in the past.) The QA lab that I worked in tested both the hardware and software for each operating system.

In 2001 they decided to support Linux, because customers were asking for it. When they asked, "What distro do you want us to support?" the answer was "Linux". The company decided to only support Red Hat, because they couldn't afford to test their products with a zillion distros. (Today they'd probably drop Unixware and support two Linux distros.)

The reality is that, due to the complexity of testing hardware/software that "has to work", each distro needs to be treated as its own operating system. The company that I worked for would probably gladly support the distro of choice from large customers, but a small company wouldn't be able to get much support for "Joe's uberk00l Linux d1str0". Even if the company were to provide source code, there's too much testing (and development) involved.

one word -- solvs the problem

[bobs666]

For the Venders I have one word.

Configure

Yes Apache Perl and so many others are portable accross,
so many if not, all of the Linux distros.
Yes just require the right librarys and off you go.

If so many OSS projects can do it for free,
why can't your vender?

I smell FUD!
I shoot my crooked arrow...

I find it's the other way around

[m874t232]

Installing a server using SuSE, Redhat, or Ubuntu is much simpler than using Windows; with the Linux systems, after you select what you want to run, it installs and just runs, and it comes with excellent GUIs for configuring it.

In contrast, getting a real-world server set up on Windows requires hours of third party software installation and configuration until it all fits together, if it ever does.

Re:Open source is easier, if you know how to work

[qray]

I do know but your average user does not. For me in both cases the alternatives were cheaper than open source. My time is valuable and a few hours of milling around news groups, googling, and IRC chat sessions costs me more than the commercial alternatives.

And documentation that is incorrect (and I'm not talking just typos and wrong names), which was the case with my Linux distro, is worse than no documentation at all. It lead me to believe I wouldn't need to go searching news groups, and hanging out in IRC and such. What was rather commical is that I had 4 sources of information and none of them agreed. I ended up piecing information together from all of them and a fair bit of trial and error. While I did get my kernel recompiled, hitting this same scenario with Postgre SQL caused me to abandon the project. Again this was many years ago and I would expect things have improved somewhat.

I hope to revist and play with some of the current distro and see how they fair. I've found the VM image distrobutions and interesting way to get ones feat wet without having to go through all the pains of setting up a full system.
--
Q

Re:Open source is easier, if you know how to work

[qray]

Whoever did the install for you fscked up, as the docu for the installed version should be available

That's quite possible. This was a few years ago, as I said I hope such things have been fixed since then. Especially since this distro wasn't free. I will say the Apache web server was one of the few things this distro had working correctly. But that was the only thing that I didn't have problems with. I need to get a recent version and take Linux for another test drive.
--
Q

Repeatable process?

[leonbrooks]

Well, no, in Linux we don't constantly get a stream of security invasions and layer upon layer of code dedicated to the prevention of this.

For example, it's technically impossible to secure display-function invokations in 'Doze, but it works safely in Linux. Is this absence of risk supposed to be a dis-advantage, somehow?

Well... only if one is looking for control over one's customers as the big aim, rather than indepoendence and safety. So... MS's grand aims still haven't materially changed all that much over the years, have they?

Linux isn't that complex either.

[einhverfr]

One of my customers has a mixed environment with Novel and Red Hat Linux systems. Today one of their employees asked me a question about Ubuntu (which, I am ashamed to say, I haven't played around with much). I was able to answer the question because I understood the system and conventions.

I would submit that it is nearly impossible to understand Windows to the same degree because of the added complexity. What was it Linus said? "Nobody understands NT?"

This being said, I have run into situations regarding open source solutions where attempts to reduce complexity seriously backfired. Try installing SQL-Ledger on Spikesource's stack, for example... And to make matter worse, imagine your Linux distro (SLES 10.0) doesn't come with a TeX distro and you have to install TeTeX from source (because you don't want to download the entire TeXLive CD). This combination is particularly difficult.

Not that bad

[einhverfr]

In the vast majority of cases, you are just a compile away from supporting another Linux distro. Sure you might need to compile a few of the more obscure libraries, but the number of those that are not distributed on most distros is dwindling.

The recompile is necessary usually due to versioning problems.

The problems occur when you try to find ways around this, such as using a standard stack that replaces many core aspects of the system. This usually adds far more complexity than it removes.

I have to agree -- solution A

[Tablizer]

I tried to get SqLite to work with 2 different version of PHP on two different machines and still haven't succeeded after long nights with php.ini re-editing and downloading and moving different versions of DLL's up, down, and all around. Perhaps some of the problem is with the Windows OS, but either way, it sometimes makes pre-packaged boxes of vendor software look inviting. Sometimes it appears that the OSS software coders coded for performance instead of ease of installation and integration.

I think OSS needs something like ODBC (but not just for DB's) where a service is registered and listable. Different products only have to be able to be registered as a service to communicate. This reduces the need to make connectors for say 10 different databases to 10 different languages, with 10 different versions each, etc. OSS seems to have a combinatorial explosion of component combinations.

When you need it...

[einhverfr]

Interesting, I've never had need to talk to a MS-SQL or any database developer to use any of the database products I've used. I've worked on some pretty complex databases and queries as well.

I have needed this, and it usuall happens when you really don't expect it.

Here is an example: I developed a custom POS solution for one customer based on SQL-Ledger. It handled a much larger number of transactions than most SQL-Ledger installations and after about six months a very odd performance problem surfaced. Suddenly pages were taking upwards of 30 seconds to load. I isolated the query that was causing the problem did an explain analyze and wrote to the email list.

Tom Lane, one of the main developers of PostgreSQL, provided me with a very clear and detailed description of what was wrong. I was able to resolve the issue in part based on his ability to tell me exactly how the planner was going about designing a query plan.

Re:Open source is easier, if you know how to work

[einhverfr]

I have worked with PostgreSQL since 6.5

Yes, in the beginning, it was a bit of a beast to work with though it was better from a data protection viewpoint than MySQL. Today, it is far and a way better than anything else I have worked with.

Re:Open source is easier, if you know how to work

[Eivind Eklund]

I agree: I've been burned by incorrect documentation, and it utterly sucks. I've also abandoned Linux, partially due to sucky documentation - instead, I use FreeBSD (disclosure: and I'm a FreeBSD developer).

And my experience with open source vs commercial is that it's usually cheaper to set up something commercial, at least as long as things go without problems - yet it's hell to debug, so if there's trouble, I'd much rather be on an open source solution. I tend to say it as "open source costs a trifle more for everyday things, and that's insurance against things suddenly going horribly wrong".

With what I do, it's usually cheaper to spend a little more at predictable times instead of taking the risk of suddenly spending lots of time when things have gone wrong and something's not working. I don't know if that's true in your situation or not.

Eivind.

That is scary convenient.

[jotaeleemeese]

Double clicking in a file with a caption "prog_name installer" which action is to launch your emerge command is much more user friendly (because it appeals to our visual instincts as a primate).

Or at lease the user should be met half way by providing a script alias names "install_software" or something of the sort.

emerge may have geek appeal but is completely meaningless to a newbie.

Re:That is scary convenient.

[cp.tar]

For some reason, it seems that when one uses the term user friendly, they really mean idiot friendly.

I, however, don't.

To (ab)use the ephemeral car analogy once more, people need a certain amount of training before they are allowed to drive.
The only reason computer users haven't needed formal training is because situations where one's ignorance will cause deaths are rather rare and limited; however, as the danger of money loss is greater than any other, some effort is made towards the training of the masses.

In the meantime, however, interfaces have been developed which allow users to master certain basic tasks relatively quickly.
Of course, certain less basic tasks become obfuscated, complicate or impossible. Interfaces beautifully arrange the 20% of capabilities needed by 80% of people, but ignore the other 80% of capabilities. (As a long-time Gnome-user, I have become quite peeved at the direction the project has taken recently.)

If I am to call myself a computer user, I have to be able to use its capabilities. You know, a computer user, not an $OS user. Or a $program user.

Newbies aren't users. They get to become users after a certain learning period.
I do so hope that this fact won't get forgotten.

Insightful? I don't think so.

[jotaeleemeese]

Poor documentation? Have you got a web browser and an Internet connection? Do you know Google's url? There is your documentation. Now, show me your manuals for your latest and greatest Windows OS. What? You did not get one? Pray, do not tell me. I am shocked, shocked, I am telling you. OK, show me your manual of Office? No manual neither? Gee, you can trust nobody nowadays.

I don't know in which planet you are buddy, down here on planet Earth most software companies stopped providing proper documentation ages ago, and very often when hey do they are written in Engrish. Or they refer you to their website.

As in poor support do not joke with us in jest buddy, unless you are paying you get squat. And that is as true for your WIndows, OS X or commerical Linux distribution.

If you are wise enough to use Linux for your bussiness then you will pay support, like you do when you use any other software infrastructure. The advantage is that you have choice of who procvides that support.

With commercial options there is only one game in town, and you drink it or spill it.

Re:Open source is easier, if you know how to work

[yuna49]

How about some examples? MSSQL is perhaps one of the easiest RDBMS to install/use of them all (yes, I've used Oracle, MSSQL, MySQL, and PostgreSQL) even if it doesn't scale as high. Put in the install CD and hit the various Next/OK buttons and you are done with a system that, by default, is pretty good.

And this is easier than:
yum install postgresql*
because....?

Hypocrites...

[Anonymous Coward]

IBS Synergy's website runs Linux. So I suppose it's not so complex after all....

Please address replies

[Anonymous Coward]

twitter, please post your answer to the two replies you received. Let's see if your "Windoze" and "stuff on Linux just works" blabber holds up as well as you think it does. C'mon, let's see it. Or is this the extent of your "evangelization" effort? To spew flamebait and then just slither away when someone calls you on your bullshit?