Slashdot Mirror
Linux Distributors Work Towards Desktop Standards
WebHostingGuy wrote to mention an MSNBC article discussing a move by several Linux distributors to standardize on a set of components for desktop versions of the operating system. From the article: "The standard created by the Free Standards Group should make it easier for developers to write applications that will work on Linux versions from different distributors. Linux has a firm foothold as an operating system for servers -- it's popular for hosting Web sites, for instance -- but has only a few percent of the desktop market."
After the talk there will be 2 Major Faction. While one may win. The Second one will go Screw you and make their own design in-spite of the the talks. That is the problem with Ego Driven Software vs. Profit driven. While they both have their advantages and disadvantage. Ego Driven Software while the Code my be better quality but have a much harder time agreeing with other people. But Profit driven Software tends to be more consistent but software quality tends to be a little lower.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
interesting that msn bills move as 'making the operating system compete better with windows' instead of 'making it easier for developers to write applications that work on different flavors.'
i would think the former is a result of the latter, instead of the other way around.
When you recognize love in another and realize how precious it is, everything else seems so insignificant.
It is an issue of consistency. If I am running Gnome I know when I am running a KDE app because it looks a feels slightly off. The same if I am using a straight X11 App. Linux for the desktop is not about Window Managers. It is about giving Developers tools to make their Apps Desktop Friendly, And ability to make sure Linux Apps look good no matter what WM you are using.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Having 921034 options to choose from is sometimes a good thing, but sometimes you have the feeling: why don't they just work all on 1 fantastic piece of software?
Because the worlds open source developers are not a giant slave pool designed to do your bidding.
Open source will always be chaotic and involve a great deal of duplication because that's that nature of the beast. The gain you get from that cost is much more open software that's developed rapidly and tends to work as a free market for ideas: the better ideas eventually win out (though that may take some time). If you want something different then you want Apple or Micrsoft with their rigid top down control structure which ensures that everyone is working toward a single unified goal (as much as is possible), and all the work is directed. The upside is consistency and a unified vision, but the downside is that the whole thing is more locked up, an often slower development cycle, and a tendency to get hit with the same stupid mistakes release after release after release just because it appeals to the guy at the top.
It's a choice and you can pick the software ecology that suits your needs. Just don't go expecting one to behave like the other on your whim - there are deep fundamental philosophical divisions about how to develop software (to let it evolve from the bottom up, or direct it from the top down) that are largely irreconcilable.
Jedidiah.
Craft Beer Programming T-shirts
Well, autopackage was designed to deal with many of the problems repository based distribution has, so, I would strongly disagree with the notion that it's just "a pretty frontend around the same problem". We've put many, many times more effort into things like reliable/easy installs than making it pretty (though there is still much to do).
This problem affects any OS. You can't install Safari on MacOS X 10.1 either, if I remember correctly. It's true that Linux suffers this problem worst of all though, because there's no unified platform, and because there's no profit motive so little incentive for developers to go "the extra mile" to reduce system requirements. But it's a separate (though related) problem to how you install software.
Unfortunately, those added software libraries differ among Linux distributors, making it hard to know if an application like a word processor will function on a particular Linux computer.
What a load of rubbish...
When I read a comment like this, I have to question a) the qualifications of the article author; and/or b) their motives. Any assertions made in the article need to be critically examined and their validity questioned after such false hoohah.
why don't they just work all on 1 fantastic piece of software?
/., where the favourite probably is raw /bin/bash, which is far more powerful, but also really hard to learn.
Because there couldn't be such a thing - it's an oxymoron.
Basically, the requirements of the piece of software would be heavily contradictory - dead-easy to use, but still incredibly powerful. Few such programs exist - because they are virtually impossible to make.
Example: file managers. On the one hand, you have explorer, finder, nautilus et al, which all are at least relatively easy to use even for a newbie. Many find them far to little powerful, especially on
The same principle holds for most other software. Either you make an easily usable, or a powerful version. The powerful version will, by definition, need a lot of learning on the part of the users, and thus can't be easily usable.
When you try to unite these two conflicting requirements, the most likely outcome is one of:
1) Cluttered interface, which intimidates the newcomer
2) Clean interface, but with all powerful features hidden away from sight so the advanced user has to look for them.
3) Millions of settings in an unmanagable settings dialog, toggling the different features on and off.
Conclusion: One software normally can't be the great software - not for every single user. The shifting requirements different individuals have will without doubt make them prefer different software - and that isn't really a bad thing. If everybody ran the same software, there wouldn't be as much incitement for developing new, powerful features!
I have a really elegant proof for Fermat's last theorem. If this sig was only a bit longer...
A few percent desktop marketshare is what Macintosh has. Seems to me that the "fractured" Linux desktop is doing pretty well already.
I think GUI, despite being prevalent for quite some time, have been very, very low in the priority list of Linux developers. The community has focused more on the low-level, kernel and architecture areas, and the rest has suffered from it. IMO, GUIs in Linux have always been an afterthought, and that's the reason they suck so much (again, IMO).
This sheds light in a key problem with open-source software: developers will work in what they want to work, not necessaily in what needs to be done.
Yeah, mod me down, see if I care.
Favorite quote: "
Nothing stops someone from bundling everything a program needs in a self-contained folder. That's how most of the proprietary apps I use are packaged. Open source devs could do this with their programs too, but it would be more effort without much benefit when the distros are going to package it anyway.
Actually, it's not because it's more effort. It's because it is fundamentally a bad idea.
If you bundle everything you need into one blob for each application, then suddenly your system has installed several hundred copies of gtk, all at different versions. Obviously this is quite wasteful of space, but even that is not the real problem. This is:
A security advisory was just released for all copies of gtk before a given version.
What exactly do you do now? You don't know which of your hundreds of applications has got that code included in it. Even if you could figure it out, you now have to either rebuild all of those by hand (if you can), or go to each individual upstream developer and download an updated version from them. If you're a desktop user then you probably aren't going to get this done, so you'll be running with known security holes in some applications. If you're a sysadmin then you're probably going to find a new job.
I would say that the ability to install security updates in a reasonably painless and secure manner is one of the most fundamental tests of any distrbution method. Applications-as-self-contained-blobs fails it badly.
After the talk there will be 2 Major Faction. While one may win. The Second one will go Screw you and make their own design in-spite of the the talks.
History disagrees. While the Linux Standards Base and Freedesktop.org projects haven't solved all of the problems -- and probably aren't fully adhered to by any distribution -- they have already made a huge difference in the compatibility of Linux distributions, and I think efforts like this are exactly what we need to continue pushing interoperability forward.
I say this, by the way, as a developer who just finished developing a cross-platform, commercial, binary-only application for Linux. The app I was working on definitely pushed the limits of the interoperability, since it was an authentication system that replaced key system components, and in spite of that it went very smoothly. The differences between the half-dozen Linux distros I had to tweak the package for were very small. Actually, the more difficult issue was making things work in spite of customizations the admin may have made -- I just had to punt on that one, making the installer intentionally brittle in the face of unanticipated modifications to, for example, the X startup scripts, and then providing the admin with the ability to customize the installer to adapt to local changes.
After my experience of the last year, I wouldn't have any hesitation about developing more "normal" applications to run on multiple Linux platforms, and I expect initiatives like this one (which is from the same consortium that brought us LSB) will continue to reduce the platform differences that cause problems. I think we may even be able to get to the point where app developers may actually be able to target LSB (or whatever its successor is called) rather than having to tweak for individual distributions.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.