I am sick of everyone's smarmy afterglow about their switch to Mac after all the "terrible" experiences with PCs and Windows.
It is, however, very understandable. Users try OS X, realize some of the problems they've ben working around for years and no longer even think about are no longer problems. They get a bit crazy and try to understand why most people still use Windows and in the process can be very loquacious and annoying. It calms down after a few months or a year.
But as a network admin, I have better control and flexibilty with PCs and AD than any Mac I have handled, and I started my IT career on Macs. The latest OS for Mac is very pretty and whiz-bang, but getting integration into a predominantly Windows environment requires additional software purchases, extra configuration issues and more time/money overhead.
So you're saying your IT department standardized on solutions that locked you into one vendor, and now that users are demanding support for other vendors, your lack of foresight is biting you in the ass. Umm, maybe next time you should consider the future and flexibility as a feature so you don't have to purchase new software that handles the use case you did not consider.
Yes... you can access an smb share on Windows from a Mac, after you turn of digital signing and reduce your domain's security level.
It's called NFS. Any OS can use it. Why did you ignore the possibility of Mac or Linux or Solaris workstations when you picked a network file system?
Every Mac Lover I encounter has the same story, "I use it at home and it's so easy. I must use it in the office!" Douchebag!
Well why don't you just ask Microsoft to improve Windows. You're they're customer, surely any company you chose to do business with is responsive to your concerns as a customer, right? Oh wait, you chose to do business with an entire organization of douchbags you have repeatedly been convicted of crimes against their customers. Good choice there.
Looking at porn at home and synchronzing data from your laptop to a domain share for redundancy while having access to Group Policy management are NOT the same thing.
No they're not. Your job is to implement a solution for the latter that actually works for what your users want to do. You do realize IT is supposed to be about facilitating user needs, right?
And the next person who shows me how awesome Time Machine is has a three word answer from me: Volume Shadow Copy. Windows Server has had this feature since 2003.
Congratulations. You fundamentally misunderstood the ways in which Time Machine is innovative. I don't even use it, but I read the whitepaper. What kind of IT geek are you if you don't actually read up on new tech?
And any company worth its salt has good virus protection, spam blocking and border security in place.
What does this have to do with anything? Since when has border security stopped malware problems anyway? You seem about four to six years out of date when it comes to business security models.
Now here comes the Mac which can make use of none of those office level features.
The Mac can make use of plenty of those office level features, if you implemented a cross platform solution instead of locking yourself into one vendor. Man am I glad I haven't had to deal with vaguely incompetent IT people with Windows only skills for many years. Maybe you should take some courses at the community college or something.
5% market share does not good anti- virus make.
Maybe, maybe not. But whatever Apple has done, it works so far. Realistically, malware is not a problem for Macs at this time. In future that might change.
When there are enough of them out there, and bored German teenagers get busy, then let's talk about how secure Macs are.
(2) The IT department decides that more Macs means less dependence on IT
(3) Less dependence on IT means smaller empires for IT managers...
Guess who gets to decide what users are allowed to have on their desktops...
Both of the last two companies I worked at had two interesting things. First, IT is part of the engineering group. Second, employees (including IT) get cash bonuses and stock options based upon the company's performance. At the last company, Apple is one of only two pre-approved hardware vendors and IT likes having less work to do to keep them running. The last time we hired an IT guy, OS X experience was the #1 most important criteria due to the number of employees running it.
Firstly, IBM used to make the Thinkpad, and the pro-Thinkpad loyalty that exists there is obviously disintegrating very quickly.
Okay I can see that as interesting, although it has been three years now.
Secondly, and more interestingly to me, are the numbers. There were 24 people in the pilot program, 22 of which responded to the survey. Of those 22, a whopping 19 actually preferred to keep running OS X on their Macbook instead of Windows on their THinkpad!
Why do you find that surprising? Among security professionals I know, that is below the normal switcher rate for those that try Apple machines (in my experience). In fact, that is lower than the switcher rate among engineers coming from Linux who tried it at my last company. The loyalty of people who try OS X is fairly legendary in the press and in the geek community.
86% of a group of NEW users to OS X, given a time enough to get used to it, actually prefer PS X[sic] and the Apple hardware, to the software environment they were previously accustomed to and on their company's own developed hardware system to boot.
Considering Apple's laptops are one of the few to consistently be reviewed better and do better in hardware reliability tests by independent testing companies, that just seems to fall in line with the numbers companies like Consumer Reports have been publishing for years.
( * install the OS on the new hardware
* configure the hardware to behave like my old hardware.
* recreate my global configurations
* recreate my and user accounts
* recreate my partitions If any) ) That's actually one step if you're installing any decent linux distro
Okay, I now suspect you're trolling but I'll respond one last time. What Linux distro is "decent enough" that installing it also configures my multi-button mouse to perform my preferred custom action with the third mouse button and to ignore the fourth mouse button entirely? What Linux distro is decent enough that the installation step knows to set my printer to greyscale by default? What Linux distro is decent enough that it knows to disable the mic and webcam by default for all accounts? What distro is decent enough that as part of the installation it remakes the four user accounts I have on my old system? Seriously, each of these tasks is a separate step that requires me to manually make changes and is a separate step from just getting the OS up on the box.
( * copy my data over
* reinstall all my software that is repositories (launch the script) ) There's another set that can easily be performed by a script in one step.
Not really since I have to copy the script over in the first place. before I can run it to reinstall software and copy the data from other locations.
* manually reinstall the software not in repositories This one CAN be automated if you buy into the package management model and build debs (not too hard, but I forget the tool name as I rarely manually compile stuff these days) yourself and install them using dpkg. Then its just a matter of catching the debs over in the copy data over step.
I'm sorry but manually repackaging all the software I install in and of itself is a bigger task than reinstalling by hand and that assumes I can repackage it. I'm not at all confident that some of the binary installers I run for commercial stuff are easy to do that with, or even practical.
* manually re-register any commercial software and hope the license # still works and I have not lost it This one can easily be solved by using free software, it works better as you can see from the steps above.
Oh yeah, the old don't do the task you want at all solution. I'll get right on that.
so, 4 steps that can be reduced to 3 by using the tools available to you properly, and can be reduced to 2 by selecting software more carefully.
I don't know if you are a troll or just the epitome of a Linux apologist. You think it is practical for me to waste a huge amount of time writing custom scripts, manually repackaging all the software I install, and not using the best software for my purpose (sometimes the only software for my purpose), and then perform numerous steps every time I upgrade. And you honestly think this is a better solution than just keeping OS X as by base OS, keeping Linux in a VM, doing nothing else, and using the one-step, trivially easy even for a novice method that OS X provides. That's just sad man.
If the problem is reliance on Windows, then Linux is the solution, not an even tighter software and hardware lock-in.
The problem for IBM is being locked into Windows... or any other single solution. The answer is flexibility and making sure their software and services are cross platform so if they need to they can deploy Linux or OpenSolaris or NetBSD based appliances, or (more likely) a combination of all of the above. I seriously doubt IBM is going to become Apple's biggest customer, but it certainly makes sense for them to make sure they interoperate with OS X such that they can sell solutions to customers that include OS X systems when they make sense. It also makes sense for IBM to consider OS X for some internal uses where making Linux work might be too expensive (like running Adobe InDesign). In general IBM seems committed to moving towards Linux, but they need to keep their options open and where Linux is not appropriate, it makes sense for them to have a choice of OS X or Windows. You'll note, these were all Windows users that were comparing OS X to Windows, not to Linux.
The article says IBM is running a very small program to let some employees run OS X. Is that news?
IBM is a giant corporation that has been slowly moving more and more away from Windows internally and has a large scale move to Linux underway. It is an engineering organization, in the computing field. OS X has been rapidly gaining market share in the US and undoubtedly many of IBMs customers use it to some degree.
It would be news if IBM was not running a small program to see how well OS X works internally, especially since they use their own company as a proving ground for things they sell elsewhere. This clearly helps them create better solutions for customers that have OS X in their mixed deployments.
The article says their employees have have a very positive response to OS X, the vast majority of them preferring it over Windows. Is this news?
OS X has been positively reviewed by most users for a long, long time and compared very favorably to Windows by, well, a lot of different people and members of the press. It has been gaining install share in the US (and slowly worldwide) compared to Windows. It has been gaining market share very quickly among geeks, like here on Slashdot and in scientific fields. It would be news if most IBM users did not prefer it.
In short, this article is "news" mostly in that it just confirms what we already know, but which many Mac users are still a bit insecure. Is there any article about IBM and OS X that won't make Slashdot?
Apparently you think that Mac OS application bundles are a Panacea. That's not obviously the case.
Not at all. In fact if you look at some of my other comments in this article you'll see a nice list of the ways MacOS style application bundles fail as implemented and are inferior to package management on Linux today. What I do recognize, however, is that they have effectively solved a significant number of usability problems and if Linux on the desktop developers are serious about "keeping up with the Jonses" or just being usable, they need to solve these same problems, just as Apple still needs to solve some of the problems Linux no longer has.
Further, your implication that Linux package management is worse than Windows packages because it's worse than app bundles is silly.
I never implied that. I stated that "installing applications" on Windows is more usable for the average user, than installing applications on Linux. This is partly because MS has ignored package management which means they only have one workflow to learn, instead of three. It also means developers on Windows have created reasonable tools to make things usable and those tools are widely used. While some such tools exist for Linux, none of them are widely used and none that I know of work across a significant number of distros or remove the need for multiple workflows.
There are trade-offs in package formats, and although I agree that Linux does still have some usability issues when it comes to packaging the issue isn't as clear cut as you're making it out to be.
The specific usability problems I pointed out are very real and clear cut. I used OS X as a way to demonstrate that they are solvable problems and not an inherent problem with computing. As for tradeoffs, none of the usability problems I addresses create new problems that end users have to deal with.
As an active user of Ubuntu, I can assure you that directing other users to packages in the repositories by IM or email is the simplest use case ever - tell them the package name and they're all set.
No it isn't. IM'ing users the application itself is a simpler use case, is more understandable, and has fewer steps. In any case, this isn't about finding the simplest use case. This is about making all the use cases as simple and usable as possible. Linux does okay at getting software from a repository to the user if someone IMs the name to them. It would be simpler yet if it was easy to IM a link to them that will open the package manager to the right package for them.
You can complain about other use cases all you want, but using that one as an example is absurd.
I'm not sure what you're talking about. I addressed common use cases such as a user learning about software from a Web page, then wanting to install it on their distro of choice. A link is a lot easier, than copying and pasting the name into the package manager. Unfortunately, there is no easy way to make a link that will work for all Linux users on all distributions on all hardware, let alone for multiple platforms. That is a deficiency that can be solved. Other use cases are a great deal more problematic.
Package managers are not panacea either, especially as currently implemented. They need to be expanded to take care of many real-world workflows more easily and to solve problems that users do not encounter anymore on other platforms. People can't go off touting Ubuntu as "more usable" and "ready for mainstream users" and then stick their heads in the sand when real problems are pointed out. At this point it amazes me that Linux on the desktop progresses at all given the attitudes of most users and developers, who seem a lot more interested in emotively defending their choice of OS, instead of acknowledging areas where improvement is useful and where they could borrow ideas from other OS's. I've ben using Linux on the desktop for many years (as well as sev
If you are installing the same version on the same target hardware, there's not a big issue.
If I'm copy the same versions to the same type of hardware then it is a pain in the butt because I still have to manually take care of the commercial stuff and other stuff that installed via binary installers. I also still have to manually download and recompile all the stuff that was not packaged. If I'm migrating to new hardware, this is even more of a pain in the butt.
Also, you don't have to be there and awake, its called a bash script.
Are you drunk or something? The steps in migrating to new hardware that involve me being there and conscious, if just to launch a script include:
install the OS on the new hardware
configure the hardware to behave like my old hardware.
recreate my global configurations
recreate my and user accounts
copy my home directories over
recreate my partitions If any)
reinstall all my software that is repositories (launch the script)
manually reinstall the software not in repositories
manually re-register any commercial software and hope the license # still works and I have not lost it
copy my data over
That means I have to be there, in person to launch each and every one of those tasks and some of them require me to perform them by hand. That means if I'm going to sleep as OS X allows me, I better set my alarm to wake me up about 7 times and be prepared to spend a good hour or so in the middle of the night working.
I'm not claiming this is not doable. I've done it before. What I'm saying is for a normal user, this is horribly unusable. On Windows they can buy commercial software to do most of this automatically. On OS X, it is a couple of clicks and key presses and then it is all automated. This is a very real way in which:Linux is inferior for normal users and one of the reasons for that is because their package management does not deal well with all use cases.
This is the point in the report that I started giggling uncontrollably, embarassing myself at the coffee shop...
For anyone who has met Tom, this is an amusing image. He's a big guy who (at least when I knew him) often dressed like he was on his way to a Siouxsie and the Banshees concert.
I'm frustrated with some users' tendency to complain about usability issues that either 1.) haven't existed for years or 2.) have workable solutions that the users obviously have made no attempt to find or 3.) really are related to expert-only tasks that are properly designed to be convenient for experts (i.e. building a custom kernel).
All of the problems I described are ones normal users still encounter and I still encounter regularly.
The issue of packaging is a real one, but in practice it's also a solved one. For a good example, try to install Skype on Ubuntu by clicking the skype.com download link.
It's not a solved one for users! It isn't a solved one for me because I still have to deal with the fact that most software is not packaged well enough for Ubuntu so obviously it is not easy enough for developers to do that. It sure isn't a solved problem for attempting to install software from a DVD or from a flash drive of a network drive or a computer that already has it installed. It rarely is as easy to install Linux software from a Web page on Ubuntu as it is to do the same task on Windows. I don't call that a solved problem. I call it a problem with a partial, potential solution that has not been implemented.
I remember a friend asking me for some specialized software. I had exactly what was needed on my OS X box, which was good because the developer had stopped distributing it. I dragged it into my IM client and he ran it. That was an example of very good usability for said task. The same task on Linux is not an example of good usability, because of the packages used on Linux. If it was originally installed by the package manager and the user knows where the OS caches them and they have the same hardware you can send the package to them and then they can open it in their package manager. If it was not installed by the package manager originally, then you almost certainly aren't going to still have the installer binary around and you're out of luck. That is NOT a solved usability case on Linux. The same goes for many of the other use cases I listed for installing software.
Package managers on Linux were a great idea, but they have not grown to accommodate new use cases or use cases that other OS's have already solved. Mostly this has to do with the failure to standardize, the outdated nature of the package formats themselves. They don't deal with all the common package formats and developers routinely bypass them and go for stand alone installers like Windows 3.0 had. This is because they aren't working for what people want to do, namely installing things easily onto any distribution from a variety of different sources, including software that is not in repositories. You can say it is a solved problem all you want, but it sure as hell annoys me regularly, as do people who try to explain away the problems by blaming them on anyone and anything other than Linux and it's application installation technology.
...that doesn't mean that people who are trying to accomplish things today shouldn't use the (perfectly functional) methods that are available right now.
You mean they should not use the imperfect and less functional methods that are available today, provided they are already using Linux and don' have access to another OS.
Just to clarify, we aren't talking about most people. We're talking about developers who have decided to use cutting-edge SDKs. People who can't handle building packages from source are, almost by definition, not developers.
The thing is, a lot of people who aren't developers might want to run cutting edge software or just normal software that isn't packaged for their distro. A lot of people download beta copies of software to try it out. As for developers, sure they should be able to do it, but do they want to? Why would I want to build a program by hand when I can just use another OS where it comes more nicely packaged and where I don't have to waste my time?
I guess what I'm frustrated by is the Linux development community's tendency to excuse usability flaws by claiming they are "features" to stop users from doing things if they aren't experts or by claiming it is good to keep it hard to use to keep non-expert users off the platform out of elitism. It is possible for users to do a lot of things, but if it isn't really easy many people will go to an alternative that is, and Linux's many problems related to the small install base will persist.
Well, their main competitor Ubuntu is basically giving away the OS for free. How can RedHat expect to compete with that?
They can compete by doing the same thing. Linux is not a product to be sold. It is a carrot to be used as incentive to make a sale of support and services. There are a lot of companies that can save a lot of money by switching to Linux and getting away from paying licensing fees to MS every so many years. Redhat and Canonical and IBM can all make money offering to help these companies do that, solve problems once they do that, and provided related services. Support contracts are usually what, 20% of a large software or hardware/software purchase? 20% is a lot better than 0% which they are getting from Windows desktop installs. It also means they have a good shot at providing some of the hardware for another chunk as well as consulting fees for the migration and software development fees for helping them port their internal apps and write new ones. Don't forget you are more likely to get their server contracts as well.
But on the desktop, I find that Windows (XP) just works without any fuss.
Maybe it does, but there's a catch, it costs you money. As a home user, you may not even notice. As an enterprise business or government department, that's a big chunk of change and it more or less forces you to commit to paying it for an upgrade eventually. Basically just a few big organizations making the switch would provide all the capital needed to polish Linux on the desktop for their needs and the needs of most other businesses and of home users. After that, it's all profit and you have a lot of great reference customers and a proven track record of saving other companies money by helping them switch.
That seems to be Canonical's plan. IBM has been doing the same thing for a while and have been slowly migrating their internal machines to both gain experience and credibility, but also to save themselves money on licensing costs. Redhat seems to be targeting server uses for the most part, but whatever they say I bet they don't walk away from bidding on major contracts to move big companies to Linux desktops.
Be honest with yourself. That's not *really* linux on a phone, at least not in a way that would ever have any influence over a user switching their Desktop OS.
I very, very strongly disagree. Most users are locked into Windows on the desktop. MS has created a significant number of ways to do this, many of which are illegal. Linux on a cell phone does help to undermine MS's portion of the install base for particular uses. For example, MS locks in users with proprietary and nonstandard Web sites and applications. Linux and OS X and anything else on smart phones that doesn't conform to IE and Windows broken versions of Web technologies undermines MS's ability to keep people locked in. If your bank made it's Web based banking site compatible with the iPhone or with some Linux based competitor that uses Gecko or Webkit or Opera Mobile, that is one less reason for you to have to use Windows on your desktop.
The same thing goes for application lock-in (developers make portable code to target phones as well, which incidentally makes Linux on the desktop pretty easy for them). Ditto for e-mail protocols, word processing files, images, music files, movie files, etc. Every time Linux is deployed it undermines a little of MS's stranglehold.
you can reinstall all applications with three cli commands and a created text file.
I wish. Anyway, writing a script to reinstall stuff in the repositories is fine, assuming it still has the same name and works on my new hardware (which may or may not be true). For stuff not in repositories, that installed with a stand alone installer or I had to compile myself there is no easy solution on Linux that I know of.
as for data, no hunting, move over/etc/*/home/* and/var/* problem solved.
Yeah, it's not super hard for the most part (ignoring other partitions). It is just one more task in the list I presented and assumes I already have the network up on both machines and an easy way to transfer data between them. It also means I have to sit their waiting for each step to complete before doing the next one. I much prefer letting the OS do the work while I take an hour or two and get lunch, or better yet while I go to bed. Having to manually perform each step means I have to be there and awake. In short, it was acceptable until I discovered how much easier it is on OS X with Linux in a VM. It would be like if my next car did not have power locks with a little keyfob. It's no huge deal until you are accustomed to something better.
For Ubuntu, it's mostly an issue of websites offering downloads simply needing to offer 1.) a.deb package if there is no package in Ubuntu or
The problem with this is you still lose the ability to keep it up do date because package managers don't know where it came from (no standard internal URL for checking for updates) It is better than nothing, or even what is provided right now, but as you point out, it is less than ideal.
an apt-url link if there is a package in the Ubuntu repositories.
I'd argue that in addition to a URL that links to a repository, there should be an easy way to link to it on a Web or FTP server and package managers should handle that as well as remember the URL and auto-magically look for updates at that location.
The question is then "That solves Ubuntu, but what about the other Linux distros" - and there's no widely accepted good answer to that.
If Linux distros can't standardize and no one distro gains dominance by a large amount, this seems like it will be broken for a very long time.
The cleanest general answer is to provide a self-extracting installer (like loki-setup) and having it default to installing in the current user's home directory.
There are a number of attempts to create these, like autopackage. The problem is they aren't very convenient for developers and again aren't standard. They also amount to running a random binary whenever you want to install something (yay secure like Windows) and they don't play nice with package managers again, so no auto-updates, uninstall, etc.
As for frequently-changing development tools, the standard on all free Unix-like system is.tar.gz source archives. Developers are expected to be prepared to deal with them. I'm a developer myself, and I can handle having to type "./configure && make && make install" a couple times to get a new SDK installed - you can too.
Sure I can. Most people probably cannot though, and to tell you the truth, with Windows, Linux, and OS X in front of me, Linux is often my last choice for installing such software since it is the least convenient. Again, this seems like something that could be easily automated. Why can't most package managers handle importing a tar.gz and building it automatically?
It just seems to me Linux on the desktop is in the state of, "we made it usable if it is in the repository and the developer specifically targeted your distro. Otherwise, I hope you are willing to give up using the package manager (probably Linux's biggest feature over other OS's)." I find it very discouraging. What I do find encouraging is that you acknowledge it is an issue at least. Most Linux developers I know don't even see why it would be an issue for users.
That's part of the reason linux is more secure. Unlike the Windows prostitute model, that wants to run everything as code, from.EXE files to email attachments and.JPEGs. In linux (and unix in general) you have to take the time to use the chmod command and give it permission.
A primary tenant of good security is that making things hard to do as way of dissuading users from doing it, is a really bad idea. It leads people to bypass such broken systems for something more usable and less secure. A well designed security system is about letting users do what they want as easily as they want, but at the same time making sure that they can do it safely.
Then when you get a virus/trojan, it's your own damned fault.
Another primary tenant of good security is that finding someone to blame does not increase security. As with Windows giving users absurd amounts of OK/Cancel dialogues inevitably resulting in people clicking OK to something they shouldn't have. Just because you can blame the user, doesn't mean the security was not breached. Good security is about giving users clear and good choices, when you absolutely have to, and all the rest of the time not bothering them. If users become accustomed to having to click "OK" to make software work, or chmodding executables to make them work, then they'll consider doing so less of a security issue and be more likely to slip. Ideally, users should only be bothered when they've run an application and the application actually tries to do something risky, then they'll actually pay attention.
Seriously. When you first started using OS X, you bought a new machine that was specifically built to run that OS. Comparing that experience to trying to install Ubuntu on random hardware is absurd.
Agreed, but once installed, there are still significant issues with usability on Ubuntu. It has a few wins too, but a significant number of losses.
If you want to compare your OS X experience to anything, compare it to a Dell with Ubuntu pre-installed.
I run Ubuntu in VM on top of OS X on a Mac. The VM was nicely customized for Ubuntu and all the hardware works flawlessly, including add on hardware I just plug into the system. I'd argue this is a better hardware compatibility experience than even using a Dell with it pre-installed, for many use cases.
That said, there is a reason OS X is running as the base OS right now and Ubuntu is running in the VM and it is not hardware compatibility. It is quite simply that Ubuntu is still behind on a number of significant features. It has some wins, but most all of them work from within the VM. It has some significant losses that would make it a drawback for a base OS.
For example, VMs generally work fast enough for all my uses, but have limited performance for 3D graphics. Application availability on the two OS's mean some of the applications I need to have fast 3D graphics are available for OS X, but not Ubuntu (not Ubuntu's fault, but a reality of the market today). Every so often I buy new hardware. With OS X running as the base OS, I can simply plug a firewire cable from the old machine to the new click a check box, and everything is flawlessly copied onto the new system (including the Ubuntu VM). With Ubuntu running as the base OS, I have to install the same OS on the new hardware, recreate my global configurations and user accounts, copy my home directories over, reinstall all my software, copy my data over, try to find all my certificates and preferences and then spend the next month tweaking configurations and discovering applications I forgot and reinstalling them, sometimes hunting down activation codes an the like. The latter consumes significant time and is dull and tedious work that occasionally results in some shareware or payware going away because I lost the license or can't get it to re-accept it.
Wins for Ubuntu are things like easier installation and upgrading of OSS packages, which also tend to run better on Linux. That works just as well in a VM as out of one though, so does not motivate me one way or another. All in all, Ubuntu still has a lot of rough edges and I hesitate to recommend it to people who are not very basic users (Web, e-mail, word processing) or high power users not afraid of learning the CLI and how to edit config files.
Oh man, not these tired old arguments again. I have mod points and I was going mod this down, but I'm in a charitable mood and feel like feeding some trolls today.
I'm a desktop Linux user, but man, some of your arguments are pretty weak or disingenuous and I have to call you on them. Calling people trolls for expressing some very common opinions is also rude.
The Linux software ecosystem is rife with applications that perform the same task as their popular proprietary counterparts. Some of them aren't quite up to par (Gimp), some are roughly equivalent (OpenOffice), and some are leagues better (Firefox). There are more and more proprietary applications being ported to Linux all the time.
There are lots of Applications for Linux, but there sure aren't equivalents for everything. The overlap of applications however, tends to favor a lot of the free/OSS stuff being ported to Windows, whereas a lot of the Windows payware, freeware, and shareware never gets ported to Linux. We're talking stuff a significant number of people need to do their job or their favorite hobby. If you want to browse the Web, well Linux has plenty of good choices. If you want to layout a magazine every month, and it is not formulaic, you really are going to want Windows or OS X.
Neither Mac or Windows come with a system where you can browse from a catalog of over 10,000 applications and install any one of them instantly, for free, with the click of a mouse button.
True, package managers are a big win, in some cases. That doesn't dispel the point. It is harder to find and install much software if you're using Linux and like a normal person doing your research on the Web or in a retail store. If you happen to already know what package you want and it is in the repository, package managers are a win. They're also a win for automated updates and several other tasks. They're still not comparable to Windows, however, if you want to do something so you look through Web pages or go to the store and ask a clerk.
This hardware myth really needs to be put to rest. Linux supports a wider variety of hardware [lwn.net] than any other operating system on the planet.
That's great and all. The thing is, if you're looking to buy hardware, you're going to have a harder time finding given hardware will work with Windows and if you're a normal person who does not install their own OS, you're going to have a harder time finding a store that will sell you a system with Linux pre-installed and all the hardware/software interactions polished.
Ubuntu and many of its derivatives will ship you a copy of their OS on CD at no charge. No media fees, no shipping and handling. Free. Most of the software that you can install afterward is not at all too large to pull down via a dialup modem. Windows and OS X cost hundreds of dollars each.
Yup, this is really nice, but most people buy hardware and it comes with an OS already installed. Asking them to order a DVD from some place, wait for it to arrive and then install it and hope it works with their hardware is worse than the average user's experience with Windows.
Not sure what you mean here. On KDE- and GNOME-based distributions, a shortcut to every installed application gets put into the applications menu.
For default apps, I agree. Never had a problem with most of the stuff in the repository.
Contrast with Windows where each application goes into its own folder or a folder named after the company that distributed it. Install enough applications and the Start menu becomes large and unusable. Contrast also with Mac, where you have to dig down into a special (and also unsorted) Applications folder to find newly-installed apps.
To be fair, a lot of Linux software is installed in odd menus, is CLI only, or installs into a directory named after the company that made it too, only free/OSS softwa
The OEM Windows OS is - for all practical purposes - a one-time purchase for the life of the system.
So? It is still a real cost. In low-end purchases, it can account for 20% of the purchase price. Don't even get me started about enterprise site license costs.
-- roughly equivalent to the price of two PC games, a pair of replacement ink jet cartridges, or a month of broadband cable service.
Two of the PC games you're talking about accounts for an entire year's worth of games for most people. Ink cartridges are artificially inflated in price. Broadband is artificially inflated in price where there are local monopolies. In other places, where they have modern infrastructure and competition, broadband by itself can cost as little as $20 US.
For which the buyer gets 100% compatibility with the hardware and software the home user wants to run.
Except you don't. Neither WinXP nor Windows Vista will run 100% of the Windows software I want to run. There is no version of many software programs I use for Windows, everything from commercial software like Omnigraffle, to OSS projects like scientific data analysis packages.
Assuming your use cases are everyone's use cases is naive.
How about Adept/Synaptic/other graphical package mangers? You skip the going to the store or website altogether. I want this, this, and this, click install.
That is great for a subset of use cases. If it is OSS in the repository and you know what you want by name, that's a great option.
If (like most people) you first learn about software by reading a Web page, you'll find that it is a lot harder for most use cases. Clicking on a link in a web page is easier than, copying and pasting the name of the software from a Web page into your package manager, even if it is already there. If it isn't there, then you better be prepared for a long and difficult workflow.
Linux does have a usability win by using package managers in some cases, but in many use cases, existing package managers ignore a huge number of ways people get software and that leads to a lot of frustration.
What's confusing about the Ubuntu "Add/Remove..." tool? It's not the same as a Best Buy, but being able to browse and install (for $0) thousands of programs is sort of nice.
You're right and you're wrong at the same time. Package managers are a big win for installing and keeping core software up to date. They're also very weak and hard to use for other use cases. If software is on a flash drive, on a DVD, needs to registered/licensed, at the store, on another computer, or on a P2P network, installation is often dreadful. Also, a huge number of people find software via the Web and learn about it there, but then copying the name of the software from the browser, pasting it into the package manager and trying to find and install it can be very frustrating, especially so if it is not in the default repository. Users are accustomed to the simplicity of installing by clicking a link on a Web page, then answering a few questions that pop up for them. By comparison, every Linux distro I've used has had significant problems in this regard. If the software is commercial and needs to be paid for and registered, it is even worse.
As an example, trying out the new development snapshot of Webkit on OS X entails, going to the development Web page and clicking the download link. Then, you double click on the icon. That's it. On Ubuntu trying out the same snapshot in Konquerer entailed, going to the Web site, then going to my package manager and searching for Webkit and Konquerer, discovering what I needed was not there, searching for a dev repository to add, discovering it did not exist or was not discoverable, going back to the Website, downloading the binary snapshot, searching forums for info, finding and downloading the source for Kubuntu, using the CLI to build and compile Kubuntu with a lot of special options, and finally running it from the CLI as well. I did this once and got it to work and tried again at a later date and gave up in frustration after several hours.
Now I'm sure some people will point out that there are probably equally broken development software packages for OS X. Someone else will think, "it's the developers fault, they should have put it in a package and got someone to add it to the Ubuntu/Kununtu repositories, as well as the other major distros." At least that is what people said last time we discussed this.
Let me head those trains off at the pass. They're both true, to some extent, but both miss the point. For OS X it is really easy to make a package that is installable from a Web page and takes little work from the developer or user. For Linux, this is very difficult for a developer and even if the developer goes to all this work for every release for every package manager and they can put it in random repositories because their software is not payware, it is still a harder workflow for the end user.
If Linux distros are serious about usability for application installing, they need to not drop the ball on package management as they now do it, but they also need to account for the realities of how easy it is for developers and users work together to run a program starting at a Web page, retail store, flash drive, DVD, network drive, or friend's computer. Ubuntu (and every other Linux distro I've used) has been problematic for these use cases and telling everyone to put in more effort and stop using the Web to find software and stop buying commercial software and only use OSS is just not a practical option.
I'd say that distributions like Ubuntu is exactly as user-friendly as OS X. If you use supported hardware and don't want to customize the OS in non-supported ways, everything just works.
I disagree. First let me say, I use both systems on the desktop daily (and have both in front of me right now). I also have formal education in and have worked in the field of user interface design and usability testing for disparate systems over several years.
I agree that not having the hardware vendor polish an install for their system is a huge source of usability issues. Most users never install an OS, and if you give someone a pre-configured system with OS X or Linux, you've solved a lot of their problems already.
That being the case, however, Linux still has some significant usability issues for many many, workflows and tasks. Linux is outstandingly usable for super-power users who need/want to create highly customized and specialized workflows and are not afraid of learning new interfaces. Linux is fairly usable for a very novice user who has a very limited number of tasks and workflows (Web, e-mail, word processing, playing CDs). It still has some interface issues, but it also has a few usability wins in this regard (such as at the task of keeping this core software up to date). They obviously have not, however, done the extensive usability testing Apple does, but they've hit most of the low hanging fruit for very novice users.
Linux has a lot of usability and interface issues when it comes to in between users. People who want to add new hardware (webcam, fancy trackball, stylus, braille board, or whatever) are more likely to have usability problems and not just because of lack of drivers. People who want to install and run software for specific more advanced uses such as: video editing, audio recording/mixing, 3D and vector graphics, publishing, or most commercial software like big games and other payware, still have significant usability problems. People still have significant problems trying to perform some common, but advanced tasks: creating a restricted user account for guests, migrating an installed system to new hardware, or sending a friend some software you have installed (but which is not in the repository), or enabling more advanced user interface features.
In short I understand and agree with your point about hardware, but I disagree in general about Linux being as usable as OS X for the gamut of end user tasks. I don't think any Linux on the desktop developer invests significantly in usability testing (based upon their resulting products) and I don't think they will catch the last 20% or so of problems until they do. I don't think they've even done enough work to address some of the fairly obvious problems that you can find and correct without such testing.
Indeed. Does Herdy already have a GUI for configuring all the buttons of a multi-button mouse? And a GUI for configuring all the features supported by Synaptic touchpad drivers (that already are in kernel)?
Hardy does not have a GUI for configuring multi-button mice, however, the left, right, and scroll wheel all worked by default just fine, which is all that most normal users need. As for the trackpad, no there is not a GUI for configuring all the functions, but again it did work as expected by default and did have a GUI for configuring some common options.
In short, while the GUI is not as full featured as OS X or Windows in this regard, it is just fine and functional for novice users.
If not, users still need to edit xorg.conf, and there is still work to be done.
I don't think many reasonable people deny that there is lot of work that could improve Linux. It does, however seem to be ready for novice users and their basic tasks. It falls down for users that need more power and options, but are not so advanced that they are comfortable learning to use a CLI.
And does it have a GUI for configuring xrandr defaults on X startup, so that users (with compatible drivers,of course) can easily set multi-monitor setups (that have full 3D acceleration support, unlike with Xinerama)?
Multi-monitor support, especially driving multiple monitors with 3D features and handling sleep and suspend states that may change (close laptop, unplug monitor, take it home, plug in a different monitor, open laptop, no disaster ensues) is pretty iffy on both Linux and Windows and is sporadic at best. The only OS that seems to get this right is OS X, and only with the latest (10.5). Since this is an unusual task still (because most users don't expect it to work) I don't see this as a big problem for usability for normal Linux users.
There is a lot of work to be done on all OS's. Some things on Linux just work, or are configurable and usable via the GUI, that Windows and OS X both fail on. The main difference is users coming from another OS have expectations as to which things will work and which won't and so get upset when Linux fails at something Windows can't do, but are sometimes ecstatic when they stumble across something Linux can do easily, but that they've never even tried on Windows. With Windows being so dominant, however, it is easy to see where people view this fairly one sided.
Don't forget Mac OS X being unready. I have to do a ton of terminal-based maintenance on Macs (to the point where I've put Terminal in the Docks of many of my clients so I can get to it faster when I need to use it).
I do quite a bit of terminal based configuration on my Macs, and I used to do the same for some Macs I helped support in my old department. That said, while as a veteran CLI user I find the bash shell the easiest, fastest way to do things, I don't know of any cases where it is the only way to make a change. Between the built in VNC and automator tools, and the XML config editing application, I don't know of any tasks I have to do from the CLI.
My question for you is, what are these tasks you have to do from the CLI? My second question for you is, do you think there is a difference between administrating networks of machines (being an IT guy) and being a normal user. Most people I know using OS X in the home and administering it themselves never resort to the terminal to perform a task. Most Linux users I know (even ones who really try) end up using the CLI in Linux to perform some task as a user (be it install software or get some hardware working). That includes myself.
I have no problem recommending either OS X or Ubuntu to the novice user who plans to use a Web browser, check e-mail, use a word processor, and that is about it. For users that actually need a general purpose computer and who plan to add hardware (a Webcam or something) or want to install software to do something in particular (play specific games or do audio editing/mixing or ray-trace graphics) I'm a lot more comfortable suggesting OS X and being fairly certain I won't have to help them eventually.
For all it's "power of UNIX" allure, OSX 10.5 is gawdawfully unstable. My Mac mini would run for weeks just fine with Tiger, but Leopard is lucky to last more than 24 hours or so.
You should really look into backing up and doing a fresh install or looking for hardware issues. Sometimes one version of an OS is less able to deal with a recurring hardware problem than a different version. For comparison, the uptime on my OS X laptop (MacBook) is pretty much however long since the last update that required a reboot. Seriously. I've had this particular machine running Leopard for 8 months or so and I've never, ever had a crash. The same was not true with the first MacBook I had, but eventually I tracked it down to a problem with the motherboard and got Apple to replace it. 10.5 has been the most stable OS I've ever run to date, including my Solaris and NetBSD servers, each of which have crashed at least once.
Do I love the Mac? Sure! Is it as stable as my Linux PCs? Not a chance. Sorry.
Let me repeat my suggestion. Really look for hardware issues. Look at the RAM, hard drive, and power supply. I've found Linux in general is more tolerant of flakey hardware, but on good hardware, OS X certainly has been as stable or more so than my Linux install, even on the same hardware.
Slashdot Mirror
I am sick of everyone's smarmy afterglow about their switch to Mac after all the "terrible" experiences with PCs and Windows.
It is, however, very understandable. Users try OS X, realize some of the problems they've ben working around for years and no longer even think about are no longer problems. They get a bit crazy and try to understand why most people still use Windows and in the process can be very loquacious and annoying. It calms down after a few months or a year.
But as a network admin, I have better control and flexibilty with PCs and AD than any Mac I have handled, and I started my IT career on Macs. The latest OS for Mac is very pretty and whiz-bang, but getting integration into a predominantly Windows environment requires additional software purchases, extra configuration issues and more time/money overhead.
So you're saying your IT department standardized on solutions that locked you into one vendor, and now that users are demanding support for other vendors, your lack of foresight is biting you in the ass. Umm, maybe next time you should consider the future and flexibility as a feature so you don't have to purchase new software that handles the use case you did not consider.
Yes... you can access an smb share on Windows from a Mac, after you turn of digital signing and reduce your domain's security level.
It's called NFS. Any OS can use it. Why did you ignore the possibility of Mac or Linux or Solaris workstations when you picked a network file system?
Every Mac Lover I encounter has the same story, "I use it at home and it's so easy. I must use it in the office!" Douchebag!
Well why don't you just ask Microsoft to improve Windows. You're they're customer, surely any company you chose to do business with is responsive to your concerns as a customer, right? Oh wait, you chose to do business with an entire organization of douchbags you have repeatedly been convicted of crimes against their customers. Good choice there.
Looking at porn at home and synchronzing data from your laptop to a domain share for redundancy while having access to Group Policy management are NOT the same thing.
No they're not. Your job is to implement a solution for the latter that actually works for what your users want to do. You do realize IT is supposed to be about facilitating user needs, right?
And the next person who shows me how awesome Time Machine is has a three word answer from me: Volume Shadow Copy. Windows Server has had this feature since 2003.
Congratulations. You fundamentally misunderstood the ways in which Time Machine is innovative. I don't even use it, but I read the whitepaper. What kind of IT geek are you if you don't actually read up on new tech?
And any company worth its salt has good virus protection, spam blocking and border security in place.
What does this have to do with anything? Since when has border security stopped malware problems anyway? You seem about four to six years out of date when it comes to business security models.
Now here comes the Mac which can make use of none of those office level features.
The Mac can make use of plenty of those office level features, if you implemented a cross platform solution instead of locking yourself into one vendor. Man am I glad I haven't had to deal with vaguely incompetent IT people with Windows only skills for many years. Maybe you should take some courses at the community college or something.
5% market share does not good anti- virus make.
Maybe, maybe not. But whatever Apple has done, it works so far. Realistically, malware is not a problem for Macs at this time. In future that might change.
When there are enough of them out there, and bored German teenagers get busy, then let's talk about how secure Macs are.
Both of the last two companies I worked at had two interesting things. First, IT is part of the engineering group. Second, employees (including IT) get cash bonuses and stock options based upon the company's performance. At the last company, Apple is one of only two pre-approved hardware vendors and IT likes having less work to do to keep them running. The last time we hired an IT guy, OS X experience was the #1 most important criteria due to the number of employees running it.
Okay I can see that as interesting, although it has been three years now.
Secondly, and more interestingly to me, are the numbers. There were 24 people in the pilot program, 22 of which responded to the survey. Of those 22, a whopping 19 actually preferred to keep running OS X on their Macbook instead of Windows on their THinkpad!Why do you find that surprising? Among security professionals I know, that is below the normal switcher rate for those that try Apple machines (in my experience). In fact, that is lower than the switcher rate among engineers coming from Linux who tried it at my last company. The loyalty of people who try OS X is fairly legendary in the press and in the geek community.
86% of a group of NEW users to OS X, given a time enough to get used to it, actually prefer PS X[sic] and the Apple hardware, to the software environment they were previously accustomed to and on their company's own developed hardware system to boot.Considering Apple's laptops are one of the few to consistently be reviewed better and do better in hardware reliability tests by independent testing companies, that just seems to fall in line with the numbers companies like Consumer Reports have been publishing for years.
Okay, I now suspect you're trolling but I'll respond one last time. What Linux distro is "decent enough" that installing it also configures my multi-button mouse to perform my preferred custom action with the third mouse button and to ignore the fourth mouse button entirely? What Linux distro is decent enough that the installation step knows to set my printer to greyscale by default? What Linux distro is decent enough that it knows to disable the mic and webcam by default for all accounts? What distro is decent enough that as part of the installation it remakes the four user accounts I have on my old system? Seriously, each of these tasks is a separate step that requires me to manually make changes and is a separate step from just getting the OS up on the box.
( * copy my data over * reinstall all my software that is repositories (launch the script) ) There's another set that can easily be performed by a script in one step.Not really since I have to copy the script over in the first place. before I can run it to reinstall software and copy the data from other locations.
* manually reinstall the software not in repositories This one CAN be automated if you buy into the package management model and build debs (not too hard, but I forget the tool name as I rarely manually compile stuff these days) yourself and install them using dpkg. Then its just a matter of catching the debs over in the copy data over step.I'm sorry but manually repackaging all the software I install in and of itself is a bigger task than reinstalling by hand and that assumes I can repackage it. I'm not at all confident that some of the binary installers I run for commercial stuff are easy to do that with, or even practical.
* manually re-register any commercial software and hope the license # still works and I have not lost it This one can easily be solved by using free software, it works better as you can see from the steps above.Oh yeah, the old don't do the task you want at all solution. I'll get right on that.
so, 4 steps that can be reduced to 3 by using the tools available to you properly, and can be reduced to 2 by selecting software more carefully.I don't know if you are a troll or just the epitome of a Linux apologist. You think it is practical for me to waste a huge amount of time writing custom scripts, manually repackaging all the software I install, and not using the best software for my purpose (sometimes the only software for my purpose), and then perform numerous steps every time I upgrade. And you honestly think this is a better solution than just keeping OS X as by base OS, keeping Linux in a VM, doing nothing else, and using the one-step, trivially easy even for a novice method that OS X provides. That's just sad man.
P.S. learn to use the quote tag.
The problem for IBM is being locked into Windows... or any other single solution. The answer is flexibility and making sure their software and services are cross platform so if they need to they can deploy Linux or OpenSolaris or NetBSD based appliances, or (more likely) a combination of all of the above. I seriously doubt IBM is going to become Apple's biggest customer, but it certainly makes sense for them to make sure they interoperate with OS X such that they can sell solutions to customers that include OS X systems when they make sense. It also makes sense for IBM to consider OS X for some internal uses where making Linux work might be too expensive (like running Adobe InDesign). In general IBM seems committed to moving towards Linux, but they need to keep their options open and where Linux is not appropriate, it makes sense for them to have a choice of OS X or Windows. You'll note, these were all Windows users that were comparing OS X to Windows, not to Linux.
The article says IBM is running a very small program to let some employees run OS X. Is that news?
IBM is a giant corporation that has been slowly moving more and more away from Windows internally and has a large scale move to Linux underway. It is an engineering organization, in the computing field. OS X has been rapidly gaining market share in the US and undoubtedly many of IBMs customers use it to some degree.
It would be news if IBM was not running a small program to see how well OS X works internally, especially since they use their own company as a proving ground for things they sell elsewhere. This clearly helps them create better solutions for customers that have OS X in their mixed deployments.
The article says their employees have have a very positive response to OS X, the vast majority of them preferring it over Windows. Is this news?
OS X has been positively reviewed by most users for a long, long time and compared very favorably to Windows by, well, a lot of different people and members of the press. It has been gaining install share in the US (and slowly worldwide) compared to Windows. It has been gaining market share very quickly among geeks, like here on Slashdot and in scientific fields. It would be news if most IBM users did not prefer it.
In short, this article is "news" mostly in that it just confirms what we already know, but which many Mac users are still a bit insecure. Is there any article about IBM and OS X that won't make Slashdot?
Apparently you think that Mac OS application bundles are a Panacea. That's not obviously the case.
Not at all. In fact if you look at some of my other comments in this article you'll see a nice list of the ways MacOS style application bundles fail as implemented and are inferior to package management on Linux today. What I do recognize, however, is that they have effectively solved a significant number of usability problems and if Linux on the desktop developers are serious about "keeping up with the Jonses" or just being usable, they need to solve these same problems, just as Apple still needs to solve some of the problems Linux no longer has.
Further, your implication that Linux package management is worse than Windows packages because it's worse than app bundles is silly.
I never implied that. I stated that "installing applications" on Windows is more usable for the average user, than installing applications on Linux. This is partly because MS has ignored package management which means they only have one workflow to learn, instead of three. It also means developers on Windows have created reasonable tools to make things usable and those tools are widely used. While some such tools exist for Linux, none of them are widely used and none that I know of work across a significant number of distros or remove the need for multiple workflows.
There are trade-offs in package formats, and although I agree that Linux does still have some usability issues when it comes to packaging the issue isn't as clear cut as you're making it out to be.
The specific usability problems I pointed out are very real and clear cut. I used OS X as a way to demonstrate that they are solvable problems and not an inherent problem with computing. As for tradeoffs, none of the usability problems I addresses create new problems that end users have to deal with.
As an active user of Ubuntu, I can assure you that directing other users to packages in the repositories by IM or email is the simplest use case ever - tell them the package name and they're all set.
No it isn't. IM'ing users the application itself is a simpler use case, is more understandable, and has fewer steps. In any case, this isn't about finding the simplest use case. This is about making all the use cases as simple and usable as possible. Linux does okay at getting software from a repository to the user if someone IMs the name to them. It would be simpler yet if it was easy to IM a link to them that will open the package manager to the right package for them.
You can complain about other use cases all you want, but using that one as an example is absurd.
I'm not sure what you're talking about. I addressed common use cases such as a user learning about software from a Web page, then wanting to install it on their distro of choice. A link is a lot easier, than copying and pasting the name into the package manager. Unfortunately, there is no easy way to make a link that will work for all Linux users on all distributions on all hardware, let alone for multiple platforms. That is a deficiency that can be solved. Other use cases are a great deal more problematic.
Package managers are not panacea either, especially as currently implemented. They need to be expanded to take care of many real-world workflows more easily and to solve problems that users do not encounter anymore on other platforms. People can't go off touting Ubuntu as "more usable" and "ready for mainstream users" and then stick their heads in the sand when real problems are pointed out. At this point it amazes me that Linux on the desktop progresses at all given the attitudes of most users and developers, who seem a lot more interested in emotively defending their choice of OS, instead of acknowledging areas where improvement is useful and where they could borrow ideas from other OS's. I've ben using Linux on the desktop for many years (as well as sev
If I'm copy the same versions to the same type of hardware then it is a pain in the butt because I still have to manually take care of the commercial stuff and other stuff that installed via binary installers. I also still have to manually download and recompile all the stuff that was not packaged. If I'm migrating to new hardware, this is even more of a pain in the butt.
Also, you don't have to be there and awake, its called a bash script.Are you drunk or something? The steps in migrating to new hardware that involve me being there and conscious, if just to launch a script include:
That means I have to be there, in person to launch each and every one of those tasks and some of them require me to perform them by hand. That means if I'm going to sleep as OS X allows me, I better set my alarm to wake me up about 7 times and be prepared to spend a good hour or so in the middle of the night working.
I'm not claiming this is not doable. I've done it before. What I'm saying is for a normal user, this is horribly unusable. On Windows they can buy commercial software to do most of this automatically. On OS X, it is a couple of clicks and key presses and then it is all automated. This is a very real way in which :Linux is inferior for normal users and one of the reasons for that is because their package management does not deal well with all use cases.
For anyone who has met Tom, this is an amusing image. He's a big guy who (at least when I knew him) often dressed like he was on his way to a Siouxsie and the Banshees concert.
All of the problems I described are ones normal users still encounter and I still encounter regularly.
The issue of packaging is a real one, but in practice it's also a solved one. For a good example, try to install Skype on Ubuntu by clicking the skype.com download link.It's not a solved one for users! It isn't a solved one for me because I still have to deal with the fact that most software is not packaged well enough for Ubuntu so obviously it is not easy enough for developers to do that. It sure isn't a solved problem for attempting to install software from a DVD or from a flash drive of a network drive or a computer that already has it installed. It rarely is as easy to install Linux software from a Web page on Ubuntu as it is to do the same task on Windows. I don't call that a solved problem. I call it a problem with a partial, potential solution that has not been implemented.
I remember a friend asking me for some specialized software. I had exactly what was needed on my OS X box, which was good because the developer had stopped distributing it. I dragged it into my IM client and he ran it. That was an example of very good usability for said task. The same task on Linux is not an example of good usability, because of the packages used on Linux. If it was originally installed by the package manager and the user knows where the OS caches them and they have the same hardware you can send the package to them and then they can open it in their package manager. If it was not installed by the package manager originally, then you almost certainly aren't going to still have the installer binary around and you're out of luck. That is NOT a solved usability case on Linux. The same goes for many of the other use cases I listed for installing software.
Package managers on Linux were a great idea, but they have not grown to accommodate new use cases or use cases that other OS's have already solved. Mostly this has to do with the failure to standardize, the outdated nature of the package formats themselves. They don't deal with all the common package formats and developers routinely bypass them and go for stand alone installers like Windows 3.0 had. This is because they aren't working for what people want to do, namely installing things easily onto any distribution from a variety of different sources, including software that is not in repositories. You can say it is a solved problem all you want, but it sure as hell annoys me regularly, as do people who try to explain away the problems by blaming them on anyone and anything other than Linux and it's application installation technology.
...that doesn't mean that people who are trying to accomplish things today shouldn't use the (perfectly functional) methods that are available right now.You mean they should not use the imperfect and less functional methods that are available today, provided they are already using Linux and don' have access to another OS.
Just to clarify, we aren't talking about most people. We're talking about developers who have decided to use cutting-edge SDKs. People who can't handle building packages from source are, almost by definition, not developers.The thing is, a lot of people who aren't developers might want to run cutting edge software or just normal software that isn't packaged for their distro. A lot of people download beta copies of software to try it out. As for developers, sure they should be able to do it, but do they want to? Why would I want to build a program by hand when I can just use another OS where it comes more nicely packaged and where I don't have to waste my time?
I guess what I'm frustrated by is the Linux development community's tendency to excuse usability flaws by claiming they are "features" to stop users from doing things if they aren't experts or by claiming it is good to keep it hard to use to keep non-expert users off the platform out of elitism. It is possible for users to do a lot of things, but if it isn't really easy many people will go to an alternative that is, and Linux's many problems related to the small install base will persist.
They can compete by doing the same thing. Linux is not a product to be sold. It is a carrot to be used as incentive to make a sale of support and services. There are a lot of companies that can save a lot of money by switching to Linux and getting away from paying licensing fees to MS every so many years. Redhat and Canonical and IBM can all make money offering to help these companies do that, solve problems once they do that, and provided related services. Support contracts are usually what, 20% of a large software or hardware/software purchase? 20% is a lot better than 0% which they are getting from Windows desktop installs. It also means they have a good shot at providing some of the hardware for another chunk as well as consulting fees for the migration and software development fees for helping them port their internal apps and write new ones. Don't forget you are more likely to get their server contracts as well.
But on the desktop, I find that Windows (XP) just works without any fuss.Maybe it does, but there's a catch, it costs you money. As a home user, you may not even notice. As an enterprise business or government department, that's a big chunk of change and it more or less forces you to commit to paying it for an upgrade eventually. Basically just a few big organizations making the switch would provide all the capital needed to polish Linux on the desktop for their needs and the needs of most other businesses and of home users. After that, it's all profit and you have a lot of great reference customers and a proven track record of saving other companies money by helping them switch.
That seems to be Canonical's plan. IBM has been doing the same thing for a while and have been slowly migrating their internal machines to both gain experience and credibility, but also to save themselves money on licensing costs. Redhat seems to be targeting server uses for the most part, but whatever they say I bet they don't walk away from bidding on major contracts to move big companies to Linux desktops.
I very, very strongly disagree. Most users are locked into Windows on the desktop. MS has created a significant number of ways to do this, many of which are illegal. Linux on a cell phone does help to undermine MS's portion of the install base for particular uses. For example, MS locks in users with proprietary and nonstandard Web sites and applications. Linux and OS X and anything else on smart phones that doesn't conform to IE and Windows broken versions of Web technologies undermines MS's ability to keep people locked in. If your bank made it's Web based banking site compatible with the iPhone or with some Linux based competitor that uses Gecko or Webkit or Opera Mobile, that is one less reason for you to have to use Windows on your desktop.
The same thing goes for application lock-in (developers make portable code to target phones as well, which incidentally makes Linux on the desktop pretty easy for them). Ditto for e-mail protocols, word processing files, images, music files, movie files, etc. Every time Linux is deployed it undermines a little of MS's stranglehold.
I wish. Anyway, writing a script to reinstall stuff in the repositories is fine, assuming it still has the same name and works on my new hardware (which may or may not be true). For stuff not in repositories, that installed with a stand alone installer or I had to compile myself there is no easy solution on Linux that I know of.
as for data, no hunting, move overYeah, it's not super hard for the most part (ignoring other partitions). It is just one more task in the list I presented and assumes I already have the network up on both machines and an easy way to transfer data between them. It also means I have to sit their waiting for each step to complete before doing the next one. I much prefer letting the OS do the work while I take an hour or two and get lunch, or better yet while I go to bed. Having to manually perform each step means I have to be there and awake. In short, it was acceptable until I discovered how much easier it is on OS X with Linux in a VM. It would be like if my next car did not have power locks with a little keyfob. It's no huge deal until you are accustomed to something better.
The problem with this is you still lose the ability to keep it up do date because package managers don't know where it came from (no standard internal URL for checking for updates) It is better than nothing, or even what is provided right now, but as you point out, it is less than ideal.
an apt-url link if there is a package in the Ubuntu repositories.I'd argue that in addition to a URL that links to a repository, there should be an easy way to link to it on a Web or FTP server and package managers should handle that as well as remember the URL and auto-magically look for updates at that location.
The question is then "That solves Ubuntu, but what about the other Linux distros" - and there's no widely accepted good answer to that.If Linux distros can't standardize and no one distro gains dominance by a large amount, this seems like it will be broken for a very long time.
The cleanest general answer is to provide a self-extracting installer (like loki-setup) and having it default to installing in the current user's home directory.There are a number of attempts to create these, like autopackage. The problem is they aren't very convenient for developers and again aren't standard. They also amount to running a random binary whenever you want to install something (yay secure like Windows) and they don't play nice with package managers again, so no auto-updates, uninstall, etc.
As for frequently-changing development tools, the standard on all free Unix-like system isSure I can. Most people probably cannot though, and to tell you the truth, with Windows, Linux, and OS X in front of me, Linux is often my last choice for installing such software since it is the least convenient. Again, this seems like something that could be easily automated. Why can't most package managers handle importing a tar.gz and building it automatically?
It just seems to me Linux on the desktop is in the state of, "we made it usable if it is in the repository and the developer specifically targeted your distro. Otherwise, I hope you are willing to give up using the package manager (probably Linux's biggest feature over other OS's)." I find it very discouraging. What I do find encouraging is that you acknowledge it is an issue at least. Most Linux developers I know don't even see why it would be an issue for users.
A primary tenant of good security is that making things hard to do as way of dissuading users from doing it, is a really bad idea. It leads people to bypass such broken systems for something more usable and less secure. A well designed security system is about letting users do what they want as easily as they want, but at the same time making sure that they can do it safely.
Then when you get a virus/trojan, it's your own damned fault.Another primary tenant of good security is that finding someone to blame does not increase security. As with Windows giving users absurd amounts of OK/Cancel dialogues inevitably resulting in people clicking OK to something they shouldn't have. Just because you can blame the user, doesn't mean the security was not breached. Good security is about giving users clear and good choices, when you absolutely have to, and all the rest of the time not bothering them. If users become accustomed to having to click "OK" to make software work, or chmodding executables to make them work, then they'll consider doing so less of a security issue and be more likely to slip. Ideally, users should only be bothered when they've run an application and the application actually tries to do something risky, then they'll actually pay attention.
Agreed, but once installed, there are still significant issues with usability on Ubuntu. It has a few wins too, but a significant number of losses.
If you want to compare your OS X experience to anything, compare it to a Dell with Ubuntu pre-installed.I run Ubuntu in VM on top of OS X on a Mac. The VM was nicely customized for Ubuntu and all the hardware works flawlessly, including add on hardware I just plug into the system. I'd argue this is a better hardware compatibility experience than even using a Dell with it pre-installed, for many use cases.
That said, there is a reason OS X is running as the base OS right now and Ubuntu is running in the VM and it is not hardware compatibility. It is quite simply that Ubuntu is still behind on a number of significant features. It has some wins, but most all of them work from within the VM. It has some significant losses that would make it a drawback for a base OS.
For example, VMs generally work fast enough for all my uses, but have limited performance for 3D graphics. Application availability on the two OS's mean some of the applications I need to have fast 3D graphics are available for OS X, but not Ubuntu (not Ubuntu's fault, but a reality of the market today). Every so often I buy new hardware. With OS X running as the base OS, I can simply plug a firewire cable from the old machine to the new click a check box, and everything is flawlessly copied onto the new system (including the Ubuntu VM). With Ubuntu running as the base OS, I have to install the same OS on the new hardware, recreate my global configurations and user accounts, copy my home directories over, reinstall all my software, copy my data over, try to find all my certificates and preferences and then spend the next month tweaking configurations and discovering applications I forgot and reinstalling them, sometimes hunting down activation codes an the like. The latter consumes significant time and is dull and tedious work that occasionally results in some shareware or payware going away because I lost the license or can't get it to re-accept it.
Wins for Ubuntu are things like easier installation and upgrading of OSS packages, which also tend to run better on Linux. That works just as well in a VM as out of one though, so does not motivate me one way or another. All in all, Ubuntu still has a lot of rough edges and I hesitate to recommend it to people who are not very basic users (Web, e-mail, word processing) or high power users not afraid of learning the CLI and how to edit config files.
Oh man, not these tired old arguments again. I have mod points and I was going mod this down, but I'm in a charitable mood and feel like feeding some trolls today.
I'm a desktop Linux user, but man, some of your arguments are pretty weak or disingenuous and I have to call you on them. Calling people trolls for expressing some very common opinions is also rude.
The Linux software ecosystem is rife with applications that perform the same task as their popular proprietary counterparts. Some of them aren't quite up to par (Gimp), some are roughly equivalent (OpenOffice), and some are leagues better (Firefox). There are more and more proprietary applications being ported to Linux all the time.
There are lots of Applications for Linux, but there sure aren't equivalents for everything. The overlap of applications however, tends to favor a lot of the free/OSS stuff being ported to Windows, whereas a lot of the Windows payware, freeware, and shareware never gets ported to Linux. We're talking stuff a significant number of people need to do their job or their favorite hobby. If you want to browse the Web, well Linux has plenty of good choices. If you want to layout a magazine every month, and it is not formulaic, you really are going to want Windows or OS X.
Neither Mac or Windows come with a system where you can browse from a catalog of over 10,000 applications and install any one of them instantly, for free, with the click of a mouse button.
True, package managers are a big win, in some cases. That doesn't dispel the point. It is harder to find and install much software if you're using Linux and like a normal person doing your research on the Web or in a retail store. If you happen to already know what package you want and it is in the repository, package managers are a win. They're also a win for automated updates and several other tasks. They're still not comparable to Windows, however, if you want to do something so you look through Web pages or go to the store and ask a clerk.
This hardware myth really needs to be put to rest. Linux supports a wider variety of hardware [lwn.net] than any other operating system on the planet.
That's great and all. The thing is, if you're looking to buy hardware, you're going to have a harder time finding given hardware will work with Windows and if you're a normal person who does not install their own OS, you're going to have a harder time finding a store that will sell you a system with Linux pre-installed and all the hardware/software interactions polished.
Ubuntu and many of its derivatives will ship you a copy of their OS on CD at no charge. No media fees, no shipping and handling. Free. Most of the software that you can install afterward is not at all too large to pull down via a dialup modem. Windows and OS X cost hundreds of dollars each.
Yup, this is really nice, but most people buy hardware and it comes with an OS already installed. Asking them to order a DVD from some place, wait for it to arrive and then install it and hope it works with their hardware is worse than the average user's experience with Windows.
Not sure what you mean here. On KDE- and GNOME-based distributions, a shortcut to every installed application gets put into the applications menu.
For default apps, I agree. Never had a problem with most of the stuff in the repository.
Contrast with Windows where each application goes into its own folder or a folder named after the company that distributed it. Install enough applications and the Start menu becomes large and unusable. Contrast also with Mac, where you have to dig down into a special (and also unsorted) Applications folder to find newly-installed apps.
To be fair, a lot of Linux software is installed in odd menus, is CLI only, or installs into a directory named after the company that made it too, only free/OSS softwa
So? It is still a real cost. In low-end purchases, it can account for 20% of the purchase price. Don't even get me started about enterprise site license costs.
-- roughly equivalent to the price of two PC games, a pair of replacement ink jet cartridges, or a month of broadband cable service.Two of the PC games you're talking about accounts for an entire year's worth of games for most people. Ink cartridges are artificially inflated in price. Broadband is artificially inflated in price where there are local monopolies. In other places, where they have modern infrastructure and competition, broadband by itself can cost as little as $20 US.
For which the buyer gets 100% compatibility with the hardware and software the home user wants to run.Except you don't. Neither WinXP nor Windows Vista will run 100% of the Windows software I want to run. There is no version of many software programs I use for Windows, everything from commercial software like Omnigraffle, to OSS projects like scientific data analysis packages.
Assuming your use cases are everyone's use cases is naive.
That is great for a subset of use cases. If it is OSS in the repository and you know what you want by name, that's a great option.
If (like most people) you first learn about software by reading a Web page, you'll find that it is a lot harder for most use cases. Clicking on a link in a web page is easier than, copying and pasting the name of the software from a Web page into your package manager, even if it is already there. If it isn't there, then you better be prepared for a long and difficult workflow.
Linux does have a usability win by using package managers in some cases, but in many use cases, existing package managers ignore a huge number of ways people get software and that leads to a lot of frustration.
You're right and you're wrong at the same time. Package managers are a big win for installing and keeping core software up to date. They're also very weak and hard to use for other use cases. If software is on a flash drive, on a DVD, needs to registered/licensed, at the store, on another computer, or on a P2P network, installation is often dreadful. Also, a huge number of people find software via the Web and learn about it there, but then copying the name of the software from the browser, pasting it into the package manager and trying to find and install it can be very frustrating, especially so if it is not in the default repository. Users are accustomed to the simplicity of installing by clicking a link on a Web page, then answering a few questions that pop up for them. By comparison, every Linux distro I've used has had significant problems in this regard. If the software is commercial and needs to be paid for and registered, it is even worse.
As an example, trying out the new development snapshot of Webkit on OS X entails, going to the development Web page and clicking the download link. Then, you double click on the icon. That's it. On Ubuntu trying out the same snapshot in Konquerer entailed, going to the Web site, then going to my package manager and searching for Webkit and Konquerer, discovering what I needed was not there, searching for a dev repository to add, discovering it did not exist or was not discoverable, going back to the Website, downloading the binary snapshot, searching forums for info, finding and downloading the source for Kubuntu, using the CLI to build and compile Kubuntu with a lot of special options, and finally running it from the CLI as well. I did this once and got it to work and tried again at a later date and gave up in frustration after several hours.
Now I'm sure some people will point out that there are probably equally broken development software packages for OS X. Someone else will think, "it's the developers fault, they should have put it in a package and got someone to add it to the Ubuntu/Kununtu repositories, as well as the other major distros." At least that is what people said last time we discussed this.
Let me head those trains off at the pass. They're both true, to some extent, but both miss the point. For OS X it is really easy to make a package that is installable from a Web page and takes little work from the developer or user. For Linux, this is very difficult for a developer and even if the developer goes to all this work for every release for every package manager and they can put it in random repositories because their software is not payware, it is still a harder workflow for the end user.
If Linux distros are serious about usability for application installing, they need to not drop the ball on package management as they now do it, but they also need to account for the realities of how easy it is for developers and users work together to run a program starting at a Web page, retail store, flash drive, DVD, network drive, or friend's computer. Ubuntu (and every other Linux distro I've used) has been problematic for these use cases and telling everyone to put in more effort and stop using the Web to find software and stop buying commercial software and only use OSS is just not a practical option.
I disagree. First let me say, I use both systems on the desktop daily (and have both in front of me right now). I also have formal education in and have worked in the field of user interface design and usability testing for disparate systems over several years.
I agree that not having the hardware vendor polish an install for their system is a huge source of usability issues. Most users never install an OS, and if you give someone a pre-configured system with OS X or Linux, you've solved a lot of their problems already.
That being the case, however, Linux still has some significant usability issues for many many, workflows and tasks. Linux is outstandingly usable for super-power users who need/want to create highly customized and specialized workflows and are not afraid of learning new interfaces. Linux is fairly usable for a very novice user who has a very limited number of tasks and workflows (Web, e-mail, word processing, playing CDs). It still has some interface issues, but it also has a few usability wins in this regard (such as at the task of keeping this core software up to date). They obviously have not, however, done the extensive usability testing Apple does, but they've hit most of the low hanging fruit for very novice users.
Linux has a lot of usability and interface issues when it comes to in between users. People who want to add new hardware (webcam, fancy trackball, stylus, braille board, or whatever) are more likely to have usability problems and not just because of lack of drivers. People who want to install and run software for specific more advanced uses such as: video editing, audio recording/mixing, 3D and vector graphics, publishing, or most commercial software like big games and other payware, still have significant usability problems. People still have significant problems trying to perform some common, but advanced tasks: creating a restricted user account for guests, migrating an installed system to new hardware, or sending a friend some software you have installed (but which is not in the repository), or enabling more advanced user interface features.
In short I understand and agree with your point about hardware, but I disagree in general about Linux being as usable as OS X for the gamut of end user tasks. I don't think any Linux on the desktop developer invests significantly in usability testing (based upon their resulting products) and I don't think they will catch the last 20% or so of problems until they do. I don't think they've even done enough work to address some of the fairly obvious problems that you can find and correct without such testing.
Hardy does not have a GUI for configuring multi-button mice, however, the left, right, and scroll wheel all worked by default just fine, which is all that most normal users need. As for the trackpad, no there is not a GUI for configuring all the functions, but again it did work as expected by default and did have a GUI for configuring some common options.
In short, while the GUI is not as full featured as OS X or Windows in this regard, it is just fine and functional for novice users.
If not, users still need to edit xorg.conf, and there is still work to be done.I don't think many reasonable people deny that there is lot of work that could improve Linux. It does, however seem to be ready for novice users and their basic tasks. It falls down for users that need more power and options, but are not so advanced that they are comfortable learning to use a CLI.
And does it have a GUI for configuring xrandr defaults on X startup, so that users (with compatible drivers,of course) can easily set multi-monitor setups (that have full 3D acceleration support, unlike with Xinerama)?Multi-monitor support, especially driving multiple monitors with 3D features and handling sleep and suspend states that may change (close laptop, unplug monitor, take it home, plug in a different monitor, open laptop, no disaster ensues) is pretty iffy on both Linux and Windows and is sporadic at best. The only OS that seems to get this right is OS X, and only with the latest (10.5). Since this is an unusual task still (because most users don't expect it to work) I don't see this as a big problem for usability for normal Linux users.
There is a lot of work to be done on all OS's. Some things on Linux just work, or are configurable and usable via the GUI, that Windows and OS X both fail on. The main difference is users coming from another OS have expectations as to which things will work and which won't and so get upset when Linux fails at something Windows can't do, but are sometimes ecstatic when they stumble across something Linux can do easily, but that they've never even tried on Windows. With Windows being so dominant, however, it is easy to see where people view this fairly one sided.
I do quite a bit of terminal based configuration on my Macs, and I used to do the same for some Macs I helped support in my old department. That said, while as a veteran CLI user I find the bash shell the easiest, fastest way to do things, I don't know of any cases where it is the only way to make a change. Between the built in VNC and automator tools, and the XML config editing application, I don't know of any tasks I have to do from the CLI.
My question for you is, what are these tasks you have to do from the CLI? My second question for you is, do you think there is a difference between administrating networks of machines (being an IT guy) and being a normal user. Most people I know using OS X in the home and administering it themselves never resort to the terminal to perform a task. Most Linux users I know (even ones who really try) end up using the CLI in Linux to perform some task as a user (be it install software or get some hardware working). That includes myself.
I have no problem recommending either OS X or Ubuntu to the novice user who plans to use a Web browser, check e-mail, use a word processor, and that is about it. For users that actually need a general purpose computer and who plan to add hardware (a Webcam or something) or want to install software to do something in particular (play specific games or do audio editing/mixing or ray-trace graphics) I'm a lot more comfortable suggesting OS X and being fairly certain I won't have to help them eventually.
You should really look into backing up and doing a fresh install or looking for hardware issues. Sometimes one version of an OS is less able to deal with a recurring hardware problem than a different version. For comparison, the uptime on my OS X laptop (MacBook) is pretty much however long since the last update that required a reboot. Seriously. I've had this particular machine running Leopard for 8 months or so and I've never, ever had a crash. The same was not true with the first MacBook I had, but eventually I tracked it down to a problem with the motherboard and got Apple to replace it. 10.5 has been the most stable OS I've ever run to date, including my Solaris and NetBSD servers, each of which have crashed at least once.
Do I love the Mac? Sure! Is it as stable as my Linux PCs? Not a chance. Sorry.Let me repeat my suggestion. Really look for hardware issues. Look at the RAM, hard drive, and power supply. I've found Linux in general is more tolerant of flakey hardware, but on good hardware, OS X certainly has been as stable or more so than my Linux install, even on the same hardware.