Slashdot Mirror
Yet Another Call for Linux Standardization
An anonymous reader writes "Newsforge has an article Commentary: United We Stand...the Division in the Linux World, in which David Meyer argues that UnitedLinux will provide standardization for the Linux community that will allow it to win the desktop market from Windows. The article has a number of supporting comments, but then this one particular negative comment that disagrees with David. This particular comment offers an alternative view on the need for standardization. This aternative view that is put forward simply argues that 'Over what is almost twelve years we have pulled ourselves up by the bootstraps. We have done this using a development model that allows us to produce software that proprietary vendors cannot compete with', and then summarizing that 'the Linux community does not need to set up businesses with the specific intention of trying to "win" users from Microsoft; all we have to do is continue to develop software in the same way, and the users will make the switch all by themselves'."
... was a enlightening /. article.
Christmas came early this year!
It's called Linus Torvalds. He will standardize as much as he can, and the rest of us will group behind the best distro of his stuff. Anything else would be closing the free developement model. UnitedLinux is trying to corner the market on useable linux.
Always going forward, 'cause we can't find reverse.
It's not lack of certain standards that makes Linux aggravating for non-Linux users. It's that those standards are so cryptic, obscure, contradictory and arbitrary. I'm not talking about TCP/IP or what have you, but simple things:
- Why is there still no standard model for adding and removing apps? The number of competing models for package management alone is sickening.
- Why do we still have to choose between a bunch of different desktops, ALL of which are mutually incompatible?
The lack of standards in Linux is even worse than the closed-ended standards on other OSes (coughWindowscough) because it makes almost any attempt to converge standards nearly impossible. We've had this for 12 years, and nothing short of wiping the slate clean is going to make it any better.
This is fine for people who don't care about such things -- who are just going to dump RedHat on a server somewhere and deal with it as little as possible. But for people who are going to be managing many different systems, not all of which are going to be homogenous, this is insanely annoying. It means that people have to learn four times as much to do the same things.
We need ONE standard desktop -- KDE, Gnome, I don't care. Pick one and use it. The others can be gravy, but we need a sanctioned interface. Not just to make things easier for end users -- and believe me, it does -- but to insure that more de facto standards do not muddy the waters any further.
And yet any discussion of such a thing in "serious" Linux circles is treated with jeering and derision. "GUIs are for wimps!" Face it -- GUIs make your life easier and anyone who tries to argue this down is blowing smoke up the wrong sphincter.
Linux users and advocates need to lose the elitism that used to preserve them, and is now working against them.
Posted as Anonymous Coward because karma can go fuck itself.
1. Unified and universal standardized library structure similar to Windows DLLs and APIs(yeah I know it's there, but it's neither standard in location or type, nor is it universal). This could also help accelerate audio and gaming library acceleration development.
/usr/local/bin. These changes are also necessary for future progression in server-side OS distros as well IMHO, but server penetration of *NIXES is (fortunately) much further along.
2. Copying the Windows registry paradigm for system and program information. One should not only be able to install programs and have their components registered, but also cleanly uninstall and/or install over existing versions in the same way. You can also standardize automatic upgrades for existing programs and kernel patches over the 'net using a similar tool.
3. GUI the hell out of every system tool there is and make sure that GUI is strictly standardized with integrated help and unified. It's getting there but it's not there yet.
4. Include copies of software with each distribution compatible to at least some extent with their Windows equivalents (e.g. XMMS, OpenOffice) though this is pretty frequent these days.
5. (Most important, and likely most difficult) Get all current developers to start working under this framework to the greatest extent possible. Whether it's open source, closed source, free software, or whatever else, a common framework is critical no matter who is developing.
That, to me, is what's essentially different between Windows and Linux on the desktop. It's a chicken-and-egg to get more developers of Windows-only software, but the only way to get them on the bandwagon is to cut a standard here and today. This is a lot more ambitious than, say, POSIX compliance. But this is what it's going to take, not just copying the binary into
From the article:
Microsoft users are an interesting lot. They have systems that they have NO control over. They have systems they have to reboot every sixteen minutes. They freely pay Bill Gates obscene amounts of money for buggy programs that they can't use when they upgrade to the next operating system. It's almost laughable. But they are united, "
Using the same OS does not make these people united any more than driving a car makes all automobile owners united.
I want to drag this out as long as possible. Bring me my protractor.
A standard does not mean that everybody is forced to do it that way. It's merely a common "language" that people agree upon. ;-)
Defining a standard will therefore enable distros to concentrate their efforts while being able to keep their own way of doing things.
Of course, if the standard lifts offs and everybody accepts it, then the distros will start dropping old features over time.
But even with a standard, it remains open source. So theoretically anybody could try to propose a new standard (as long as it is backwards compatible).
The ENIAC Demo Competition
Comment removed based on user account deletion
Because of all microsoft's bullshit, people are not used to making choices about their computers.
With opensource, you can choose what you want to do and how you want to do it, and not worry about any EULA telling you what you can and cannot do.
Not one of these statements is true (except perhaps the control over the OS statement, depending on how you define control).
I never have to reboot W2k or XP, except during the occasional (hehe) patch.
I know people that still use Office 97 on new operating systems. In fact, MS catches a lot of flack for maintaining backwards compatibility. And now we're claiming that they don't?
Microsoft users are not united. We are just customers that use the (arguably) best (or only) tool for the job (exchange, 2000 for desktop PCs, office, etc). There is basically no sense of community for MS users that I have ever stumbled across. Microsoft developers have a few hangouts, but most of us just hit MSDN when we need info.
Most (if not all) of the Microsoft users I know of (developers, admins) not only know of Linux, but have used it when appropriate. Given that UNIX is still quite pervasive, finding the robust, free version isn't that hard. Could it be, perhaps, that they only use Linux where they feel it is strong (webserver, etc) and that is the reason it isn't as popular as zealots think it should be?
As for standards... people seem to forget that Windows is top of the heap, and the Windows environment is the least standardized environment I have ever seen. Every app has to be skinnable. Every save dialog and open dialog customized beyond recognition. Just go to the Interface Hall of Shame to see what I mean.
Robots are everywhere, and they eat old people's medicine for fuel.
"Microsoft users are an interesting lot. They have systems that they have NO control over. They have systems they have to reboot every sixteen minutes. They freely pay Bill Gates obscene amounts of money for buggy programs that they can't use when they upgrade to the next operating system. It's almost laughable."
Nothing in these statements is true. Please stop using the argument that Windows is unstable (beginning with Win2K). If you are using supported hardware it's as stable as Linux and dare I say MORE stable than Linux/XWindows. (Random X crashes do occur on occasion)
Please define "NO control over". If you're talking about being able to swap VM in the kernel then yes. If you're talking about being able to choose what apps to use or themes or such than no.
My father still uses a Windows 3.0 app on his XP machine with absolutely no problems whatsoever including printing! That's one thing Microsoft has done right, being able to use most legacy apps.
I totally agree that Unification is necessary to an extent but get your facts straight before you start bashing Windows.
Eddy.WriteLinux.Com
The point you are missing is that MOST Linux developers are not selling anything. They are just developing software for their own needs.
This tends to create a system that is more developer friendly because it meets the needs of developers well. The theory is that a very developer friendly system will ultimately be a very good platform for developing any software.
I'm not sure how successful this has been, but that's what we have. Don't ask Linux developers to be salesmen, they won't like you very much. Now, there are those who are trying to sell the wares these developers have created, and it may be that they will speak for users and be able to leverage this good development platform, but I guess we'll just have to wait and see how this plays out.
So far, there's some indication that it's worked well in some areas, for example server software and appliances, and less well in others, such as desktop software.
Why? Simple. As a software vendor i would like to port my application to Linux. But what distribution should i support where it comes to libs and directory layouts? Red Hat? SuSE? Gentoo? Debian? Mandrake? Slackware? etc. etc. etc.
I have only a limited amount of time to make my product compatible with the os. If i have to support all of them i would have to make more money of my customers just to cover the costs. This would make my product not very attractive to users, and i will probably not sell enough of it to support my efforts. So i decide not to port it yet and wait for better times. The other option is to choose just one distro like so many other vendors (Red Hat anyone?). Making that distro the de-facto standard, not because of the fact that it is the best but because that is the one on which most commercial software runs.
So standardisation is good. It attracts commercial software for all distro's which will attract new users who will make Linux to be able to reach new heights.
Now, i know that OSS could compete on alot of levels with commercial software so it would not be necesary to have commercial ones but not all of them are as good as the commercial product. For alot of software there simply is no OSS alternative which could be viable. Not yet anyway. (e.g. Visio (Kivio comes close but that's it), Dreamweaver, Video-editing software (professional versions) etc. etc.)
Just like we need one type of car, one type of TV and one type of VCR.
I find it amazing that people clamor around the concept of one type of LINUX, but yet will buy a specific VCR, Refrigerator, TV, car clothes.
Why is this? Because a specific vendor has said that there should only be one user experience and not multiple. Why did this specific vendor do this? Because otherwise there MIGHT even be competition. And as a result a whole slew of minions argue along and fight into the hands of that specific company.
What we need to do is convince people that there is choice and that people can choose. Just like you can choose a VCR and TV. Interesting, is it not. You will spend hours deciding which TV you should get with the feature set, but spend one minute on the OS....
Tells you something yes?
"You can't make a race horse of a pig"
"No," said Samuel, "but you can make very fast pig"
He was referring to generic cars. He's right.
Mac users are like Mustang owners. So are Linux users, though Linux users are more like "muscle car" owners, each with their favorite version (distro).
Windows users, on the other hand, drive a Chevy Lumina, or a Ford Escort. They don't band up.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
Although I sometimes play around in Linux with my website, and whatever else I feel like wasting time with, the fact that I need help to do almost anything in Linux prevents me from using it full time.
Example: To get my ADSL connection up and running it took me well over 2 hours to find roaring penguin, install it, and configure my two network cards. In windows, I needed to click twice on the setup.exe, fill in a few forms, and that was that. In winXP I don't even need the setup.exe.
Keep in mind this is only one example. I haven't written about my printer, scanner, ugly fonts, java, and a whole lotta stuff that's taken a lot of time to find, install, and configure.
Although there are advantages to using Linux, the disadvantages have kept me a windows user, and I expect that will continue for another few years.
using System.Awesome;
Standardizing Linux is the wrong way to go about bringing Linux to the corporate desktop and the end user. But that's not saying standards are bad. Instead, the approach should be that we offer the different alternatives to what will be a standard, and then let the decision of which will be that standard for those end user be made by those end users. In other words, let the strong survive. Let there be a system that does get chosen for the new age of desktop computing, and let it be based on Linux. The semantics there is important. It should be based on Linux, not assimilate it.
Distribution choice is a good thing. But if a group of people making a few different distributions want to make changes to theirs to make sure they are the same as each other, let them. That's their choice. But corporate IT decision makes are going to be asking questions like "what is the difference between this distribution and that distribution?" So what will the answer be? Are we going to be able to say what the difference is, or will be end up confusing them more by saying "Oh, they're just alike; flip a coin to decide."
Of course, making sure that programs can be installed on, and run on, a wide range of different distributions is a good thing. But part of the responsibility to achieve that lies with the developers of that program, such as being flexible as to where files are found, what library versions can be used, etc. Consistent interfaces help, but we also need to be able to change and adapt to make things constantly improve, and when there are new things to adopt, new decisions have to be made, and choices have to be available to decide from.
Just don't move towards the notion that a single standard shall define Linux, and no other can be Linux. Linux is a class of systems that have diversity and can adapt. That is as much a part of the power of Linux as is its strength in security and reliability.
Business decisions are all too rarely made on the basis of long term planning. Regardless of the intent, those decisions will be constantly made over and over as the years go by, and as many projects fail. The needs will change, even if they are clouded by uncertainty. Linux, too, will fail, if it loses its ability to adapt.
now we need to go OSS in diesel cars
does not exist for most of us. I must disagree with the commenter's post on Newsforge about standards not being needed; Linux is already quite standardized *from a technical standpoint* eg ANSI c/c++, FHS, POSIX, sh behvior.
Of course all that depends on the individual vendor's implementation.
Linus himself did not create his kernel to compete with anything; everyone else re-created it to do that. Linus has gone on record as saying he does not really care what happens in user space; he's not interested in anything there.
Let us not forget that distro != Linux.
My next argument is that Linux distros *do* need to standardize on the UI if they want to get $LARGE-BUSINESS-ACCOUNTS. Excuse me, but have you ever tried to tell your management that they don't need to standardize? Bear in mind that in the US business place, MS *is* the standard, mainly on the desktop and 3/4ths on the back-end.... any change will probably freak them.
Leading right back into my previous paragraphs.... business management doesn't really give a crap about obscure (for them) technical standards as long as they can do their jobs effectively (again, the UI thing) which in turn puts paychecks on the table. I feel that this sucks, myself, but that's how it is, and I *do* need to pay my rent.
At the end of the day, the *real* focus of linux is a 32 and 64-bit multitasking, multiuser capable kernel licensed under the GNU GPL, with supporting libraries and tools from GNU. That's all.
Anything else is up to the rest of us.
C|N>K
IMHO, the first priority would be a group focused to define a standard set of "user data" API. .....
.userdata in every home directory having a structure something like:
_ folder>
It is amazing to see Evolution creating his own mailbox directory onto a system where the user already have thousand of e-mails into different KMail folders. It is amazing to see KAddressBook creating a empty contact database when Evolution have in his own database all your friends data.
A new project started: Chandler, and guess what??? One of the first issue on the discution lists was: We need good import filters from
What is nedded is a set of API for e-mail (sending/receiving, addressbook, notes etc....).
A set of GUI agnostic (i.e. NO GUI at all) libraries to access into a unique mode these data. If I receive a e-mail when Evolution was up, it should be available into KMail. And When I install Chandler, it must start without ANY setup required, since all my folders, pop/imap/smtp/dial-up settings are there already.
Well. A begining may be extremelly easy. A file called
<userdata>
<desktop_folder name="My Documents" type="x-application/documents">Documents</desktop
<desktop_folder name="My Music" type="x-application/audio">mp3</desktop_folder>
<mail_folder name="inbox" format="unix" type="mail/incomming">Mail/inbox</mail_folder>
<mail_folder name="outbox" format="unix" type="mail/outgoing">Mail/outbox</mail_folder>
<mail_folder name="sent-mail" format="unix" type="mail/archive">Mail/sent-mail</mail_folder>
<!-- and so on for addressbook, calendar etc... -->
</userdata>
Just my 2 cents, mtm
Sure it does: OSS exists only because of contributions.
There's not much in life for free, so if someone's dumb enough to actually give me something for free, I won't look a gift horse in the mouth.
Anybody is free to use free software. But if they want free software to work differently from the way it does, nobody has an obligation to fulfill their wishes; if nobody else volunteers, they either contribute the changes or pay for them.
- Why is there still no standard model for adding and removing apps? The number of competing models for package management alone is sickening.
- Why do we still have to choose between a bunch of different desktops, ALL of which are mutually incompatible?
1. There are many standards actually (RPM, debs, etc.). RPM, used by RedHat, Mandrake, Caldera and pretty every distributor that count beside Slack and Debian, is currently the dominant one.
2. Wrong. Desktop are actually COMPATIBLE ! You can run a Gnome application in KDE and vice-versa. Some aspect of the DE are not compatible, like themes for example, but could you use a Winamp skin in WMP ?
Another "too many choices is bad" armchair advocate trolling. Please go get a fscking clue.
:wq
Maybe he doesn't know what the hell that means?
Come on. People who are trying to migrate don't expect things like that. If you install one of the common Linux distros you're not going to be introduced to that.
Tell me how the opportunity to "restart your x server" is better than having an app shut down with a single "illegal operation" error after which your OS functions perfectly..
Free iPods - now in the UK!
1) Use 2000 forever and M$ goes out of business.
2) Upgrade (downgrade) to XP.
3) Learn to use Linux.
I know #3 will take some effort, but at this point it is almost equal to option #2. I'd put money on the prediction that, as Linux grows, M$ product quality will either get worse or the price will go up. Their shareholders aren't particularly excited about the prospects of MSN and XBox in a post-MS-monopoly marketplace.
No matter what you choose now, #3 will eventually become the best choice. Personally, I'm going to try to use Win98 forever, or at least until I can play recent games in Linux.
"I assumed blithely that there were no elves out there in the darkness"
To abuse your analogies:
=>Different TVs, but they all can view the same
=>channels and use the same antenna connectors.
PAL vs. NTSC?
=>Different VCRs but they all use the same tapes
=>and work with any TV.
Beta vs. VHS, region coding?
=>Different cars, but they all use the same gas
=>and standardised oil grades.
Regular, unleaded, diesel?
=>Differnt refridgerators, but they all use the
=>same electricity.
115V, 60 Hz vs. 220V, 50 Hz?
We're really delving into economics and economic network externalities (which have nothing to do with packets).
I recommend this as a non-technical, yet excellent analysis of WTF is going on.
The do-it-yourself spirit that has me pondering ordering 4 Lindows boxen off of www.wallmart.com and IABCOT in my basement to support some research for school simply Does Not Translate into a general prophecy that Linux will rule.
The sheep remain sheep, and will not forget that BeelzeBill is their shepherd, and they shall not want (too frequently).
Linux standards development will continue along its present, Darwinian lines. For example, we gripe about Gnome/KDE, but I haven't heard much about alternatives to X. You can say all you like about Bluecurve, but that's the general direction that things, over time, are likely to go.
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
* Free as in Freedom.
:(
Yes i can say that without ranting like a hippie. The wonder of Linux and the benefit of Linux is that there is so much more to choose from. The less standardised, the more freedom. Sure this makes applications difficult, but now we have a natural evolution... a survival of the fittest and most versatile librarys and APIs.
* Standardised Linux would attract more windows user and bill gates is satan and must die...
Ranting lunatics. Linux has a purpose, and as Free software (as in beer this time)... its performance is NOT MEASURED IN MARKETSHARE. Do not forget this. Linux should allways be for the power user, the tweaker, the guy/girl that cannot stop fiddling with their computer.
New users should be encouraged to fiddle, they should be given VMWare or VirtualPC and a nice easy distro, or they should have a dual boot system set up. Standardisation is only going to piss off the 75% of people that prefer a different standard.
Who really cares what percentage of boxen runs linux? Sure it might attract a better quality of drivers or closed source apps, but to do this by sacrificing the core Freedom values by standardising bits is ridiculous.
And all those that take marketshare as an ego thing, you are a bunch of morons. If popular equated useful we would all be running Windows.
God, reading thru my post im starting to rant like stallman
those who control the past, control the future. those who control the present, control the past.
Hey if he had to reboot to make the mouse work, then thats a valid complaint.. Nothing made clear to him that he could also have restarted this thing called an 'X Server' (whadeverdatbe). He has been told sometimes rebooting helps (windows using friends or previous experiance), so thats the only thing he can try to make it work 'magicly'.
Please try to keep that perspective in mind before you 'bitch' at 'users'. We want people to use linux? then we will get users! If something is not obvious, then we 'developers' made a mistake.
Otherwise we'll forever have linux stuck in the 'By technicians, for technicians' era.
Okay, after reading through the posts I think that there are a few things that need to be made clear.
1. Standardizing Linux distros does not mean that EVERYTHING would be the same. There is still room to customize. What standardizing would mean is that programmers would be reasonably sure that each distro would have a standard base from which to work. For example, as a developer I would know that I could count on certain libraries being available and that those libraries would be backward compatible so that I wouldn't have to recompile my products for each new version of a distro.
2. I would also like a standard way to handle copy/paste so that I know that other applications would have access to the data that gets copied from my product and that my product would have access to data copied from other developer's applications.
3. Standardizing Linux does not mean that we would only have one desktop. The most popular desktops are KDE and Gnome and clearly we already have programs that run on both desktops quite nicely. However, it would be a really good idea for these two rivals to get together and agree to standardize certain things, if possible, in an effort to make both desktops easier to support. This would be good for everyone.
A base standard for Linux distros would help developers develop their products and be assured that they would run hassle free on most flavors of Linux. This is good for the developers and good for the users.
Remember that no one is forced to comply with any standards but those who do will be making it easier for developers to support their distribution. I'm not sure that United Linux is the way to go but it wouldn't hurt to look at the standards that they intend to adopt.
The race isn't always to the swift... but that's the way to bet!
Believe it or not, there exists a breed of technically proficient users out there who fully know of linux but are windows users regardless. Like myself, I know many such people on my college campus. I have mandrake installed, and use it every once in a while. I have been using it much less since my semi-successful attempt at upgrading kde. Primarily however, I am in win2k. Contrary to popular slashdotter opinion, with some careful set up, win2k can be quite stable and secure. I am a cs major with an emphasis on AI, and I just want to be able to code. I don't have the time to learn the ins and outs of an operating system when a much more hands free one is available. As long as my OS lets me code with minimum headaches, that's what I'll use. I realize the value of linux and actively hope that it will eventually become hassle free enough to support my video card and let me install or uprgrade a program without competing standards that will work on any desktop, but that is not where it currently is. Meanwhile, win2k is out of my hair, my compiler works, and any hardware I add will also work. I can upgrade or downgrade any of my programs in a few minutes. If linux standardization fulfills its promise, the addition of my kind of user to the linux family will be invaluable.
You're right it is the apps. There is a certain amount of conceit that goes into thinking that because something is better that people will flock to it. Witness OS/2 vs Windows. OS/2 was vastly superior to Windows but who has the market share and the apps? Your average consumer doesn't care what the operating system is, they just want to run the software that they find useful. It's a small group of people that care about the operating system. How many people cared that Betamax was superior to VHS? A lot of money needs to be pumped in to Linux to bring about one common interface that is well documented and easy to code for. I think that if somebody were to really provide a real cross platform development system that they could actually target multiple platforms and if Linux gets as polished as the MacOS and Windows you might see something. If you had a 2 PC's at CompUSA, identical except for the operating system, and with all the same applications looking and working like consumers expect them too Linux might have a shot. The vast majority of consumers see a computer as an appliance and treat it as such. An example of the kind of functionality that is needed is being able to throw a blank cd-r into my drive and just drag files to the cd icon on the linux desktop to burn them. I can do this on my girlfriends iBook, I'd love to be able to do it under Linux.
Microsoft users are an interesting lot. They have systems that they have NO control over. They have systems they have to reboot every sixteen minutes. They freely pay Bill Gates obscene amounts of money for buggy programs that they can't use when they upgrade to the next operating system.
Not a single assertion in that quoted text is true. I stopped reading after that point, as someone so obviously out of touch with reality couldn't possibly have anything _useful_ to say.
As to the issue at hand... I always find it most entertaining that so many of the people who extol the benefits of standardisation for things like network protocols think standardising the OS is a bad idea. The same arguments that make standardising on something like TCP/IP a good idea also make standardising the functional basics of an OS a good idea (and if you don't consider the interface to be a piece of base OS functionality, then I think you're well and truly our of touch with the "common user").
"...I believe that if Linux is to be all that it can be the Linux world needs to UNITE behind standards."
:)
Standards are great, but I am just concerned about any "centralized" body with any sort of motives to gain control over how linux is to be made.Who the hell funds "The United Linux campaign" I don't know, it's smelling of politics...
I unfortunately can't use Linux yet cause I can't be guaranteed all my Macromedia software will run on it... will Linux standards fix this?
Is there really that severe of a division in Linux versions that if I get my box running Red Hat one week I can't make my next upgrade to SuSE?
If it's soooo bad between Linux distros then it's __got__ to be bad going from Windows to Linux... maybe that can be cleared up for me. Is it really that hard to upgrade to Linux from windows? And if __that's__ an easy switch, how hard can it be going between distros?
I really think eventually the OS distro won't mean squat when I can run all my apps in Mozilla.
Here's the crux, what would it be like if there were 10 different distros of Windows out there? And there was a standards body governing it?
-v
The users need to pick one. I picked one. They can, too. What? Are they afraid the might be the wrong one? But they aren't afraid of having us pick "the one" for them? Then they should hire one of us to pick it for them. Sheesh. Why is this so hard?
What most users want is for it to work exactly the way they are used to computers working, only better. Well, some don't care about the better part. Actually most don't give a rat's arse if it's better. They just want it to be easy and simple and do what they are doing now, which has been pretty much molded by their past with Microsoft Windows.
The real issue being raise regarding standardizing Linux isn't about what users want, anyway. It's about what developers want. It's about what lazy developers want, which is to not have to figure out anything about a different distribution. If two distributions are identical, or if there is only one, they probably don't care.
now we need to go OSS in diesel cars