ATTENTION: You are viewing a page formatted for mobile devices; to view the full web page, click HERE.

Main Area and Open Discussion > General Software Discussion

What linux needs?

(1/10) > >>

After reading the "13 reasons linux will not make it to desktop" I started this thread to discuss the problems you faced in linux & what you expect with Linux,what you think it is not making up as Desktop in Future for most of the computer users.You can post here which softwares or features missing that makes windows still a good choice.From my side i think linux misses few applications mentioned below:By the way it is my perspective so some of you might disagree.

1. Game design software (easy,effective like Game maker)
2. Theme designer for KDE,Gnome or XFCe etc.
3. GUI Builder IDE like (Borland C++ or Delphi) >> This is needed badly <<
4. Autohotkey like scripting for GUI programs & automation tasks
5. Less bloated/whistles/tabs Windows explorer like interface for Konqurer (KDE) & Nautilus(Gnome)
6. KDE & Gnome needs develops section needs lot of documentation so that new developers can step up to the project easily.

You opinion could help in Linux some way or post your views.Please post your views about it.

I don't think "game design software" is very relevant; it might be a cute thing to have, but the "real games" aren't created that way, anyway.

If linux was to be a general desktop replacement for everybody, imho what's needed is more along the lines of...

1. uniform environment (which is never going to happen). There's too many distros, too many windowing systems, too many desktop systems (GNOME/KDE/whatever). Yeah sure, you can have gnome programs run on KDE and vice-versa, but it just doesn't feel right - applications need to look consistant. At a slightly lower level, people should finally agree on a filesystem structure (not least startup scripts).

2. for end-users, some of the details have to be hidden (GNOME hides details, but it hides too much and too little). Like, regular users really shouldn't see /dev and /proc, it does more harm than good.

3. documentation. "read the source" just isn't good enough.

4. better hardware support (especially graphics, and WiFi too I guess), and less "drivers must be opensource" fascism (yeah, would be nice, but there's a lot of R&D and NDAs, and that's just the way the world works).

...and this is assuming that people can find replacements for windows software and don't need their old apps too badly, don't need too complicated file interchange with windows, etc. There's also some more tech/developer points, but those don't matter to the end-user (although they do matter wrt. "going there").

k/ubuntu have come a long way, but are by no means perfect. Having to edit the xorg config file just to get multi-mon support? And all the hoops you have (had?) to go through to get JAVA support? eek.

I love Linux and answering these questions makes my stomach feel funny, but I do it for the good of the community...

1- Game design software? What did you have in mind?
A casual peruse through Ubuntu's repositories turns up more 3D/MUD/FPS/TM engines than you can shake a stick at. I guess there's more to games than that, but that's where the money comes from...
The point is, f0dder's right. Drag 'n drop game software is fun and has its uses, but REAL games get done where the function call meets the stack pointer... ;)
2- Once again, more of these than you can shake a stick at, but none of them are 100% complete. They should be written by the guys who have their hands wet with the environment you're in (KDE tools should come from the KDE team, etc. IMO it's their responsibility...), follow the standards where they apply, and have an intuitive interface.
3- If you're using C++ and GTK+, try Glade. Also, look into Lazarus if you're into Pascal.
4- Autohotkey came about because Windows fell far short of what was needed in this area. Linux is constantly lauded for the power of it's shell scripting and many equivalents of Autohotkey's functions can be found in the average distro's /bin and /sbin directories, but I agree there's not a simple tool for writing simple scripts. Kommander looks promising, and Zenity is nice for adding handy pop-up dialogs to your shell scripts. A feature-for-feature Linux equivalent of Autohotkey would be exciting, but until then you could always learn Ruby. :)

1- You're right, and the community would go ballistic at the proposition of being 'forced' to use one look anyways. Personally, I like Windows' consistency, although the skinning community exists because they DON'T.
IMO, Differing environments/distros/window managers are only half the problem. What's happening is that Linux is always being evolved and a lot of very useful apps are written (and still being written) with ancient toolkits (xforms, tk, pre-1.0 gtk, etc.) and since it 'just works' we all live with a GUI that looks more like Windows 3.1 than KDE, Gnome, etc. I don't think a consistent interface is an impossible goal, but some things have to be set straight. The different elements which make up the 'look-n-feel' should be separate objects, each one completely responsible for it's job. The toolkit should call the theme engine, which should be part of the environment but autonomous so whichever one you prefer can respond to whatever window manager/environment you're using. For example, the toolkit says "context menu", the theme engine should say "blue with white edges, a gradient on the title block and Helvetica for the font", and the environment should say "snappable to the screen edge, and on top of the root window using the Foobar icon set."
2- I disagree, to a point. Sure, new users shouldn't have to know or even think about what the heck /dev is, but I think if there is ONE hurdle new Linux users should understand is that Linux is not Windows, My Documents/Computer/Pictures doesn't exist (there's no place like '/home'...) and it's not profitable to go snarfing around the filesystem looking for goodies. Give them a manual that tells them what everything is, skipping the details, but enough that when someones mentions it they don't panic.
3- Grrr... I HATE that. Man pages are useful IF they have real-world examples, but carefully commented code does NOT equal documentation. EVAR!!!
4- I couldn't agree more. Although the reason for the cry for open source drivers is BECAUSE of the poor Linux support most hardware companies have given. If they made decent drivers, I don't think too many folks would care...

Now for my own 2 cents.

1- Make simple things simple. I shouldn't have to hand-write a .desktop file with root privilege just to stick something in the Applications menu. Really. And I second f0dder's vote on hand editing xorg.conf. Even though I know how to do it blindfolded, I shouldn't have to jump through that hoop for 1024x768, let alone dual monitors.
2- Second vote for a consistent interface. Use whatever toolkit you want, I don't care, just let the environment's theme engine take care of the chrome.
3- Seconds on the startup scripts. Although they should never have to be touched by human hands, they should at the very least be logical and humanly understandable JUST IN CASE I need to get at them in an emergency. A GUI would be nice, but that's up to the distro to provide. I don't mind hand-editing if something is really wrong and that's the only way to go about it, and I could count on one hand the times I have HAD to do that, but it was never pleasant. Geez guys, I understand shell script kung-fu is necessary, and the comments are helpful, but don't make me paw through 17 different files just trying to figure out how my $PATH got set.
4- Consistent printing interface. Don't even get me started... >:(
My printer works beautifully, mind you, and CUPS is very good at getting things to work, but... If I want to do more than set the papersize...

I don't have time to respond in detail, but if there is one thing the Linux folk should agree on, its binary interfaces. Or even just the concept of an interface. I have no problems with the 10 million windowing systems and text editors and config files. These are the reason OSS software is so diverse and asking them to standardize on a single windowing system is ridiculous and unreasonable.

What I do want to happen is for someone like Linus to come forward and define a standard API for applications, OS and user actions. And for every distro to program to and adhere to it. Strictly. IMO thats the really hard part. Its just too easy for anyone to write an app that reads a bunch of entries from a random txt file in usr/etc/config/blahblah, makes a few calls to the window manager, and becomes a Linux app.

A good example is package management - nowdays almost every decent distro has some form of apt-get. Yet the repositories are not compatible and neither are the managers. It would be trivial for all of them to use the same format to exchange package info, but they don't. Why does Linux still suffer from dll-hell of the worst kind. Why is recompiling the kernel considered acceptable? Why, after 5 decades, can Linux/Unix still not have binary compatibility between apps.

Decide on a unified interface for your control centers and let each distro put their own fancy UI on top. Have systemwide sound, printing, clipboard etc. I don't care if its gnome/kde/qt/elf who's at fault.

Ditch the source. Every linux hacker should be made to work for a month on a system where he has no access to source code, text editors, cmd lines or compilers. As a result she'd be forced to write tools for users, not programmers.

Look at OSX. Its a shining example of how you take a powerful kernel and build apps on top of it that isolate the user from complexity without sacrificing power.

Look at Firefox. Its incredibly succesful because no one who uses it has to look at code. Extensions are 1 click. It looks good. Why can't your kernel and apps follow the same model.

Unfortunately I think the Linux community is too much in love with its own product. Ultimately, the users of Linux are mostly those who wrote it - the ultra knowledgable geeks. They will not feel the pain. They cannot, no matter how hard they try to look at from another point of view. And unless they do, it will remain what it is today.

Unfortunately I think the Linux community is too much in love with its own product. Ultimately, the users of Linux are mostly those who wrote it - the ultra knowledgable geeks. They will not feel the pain. They cannot, no matter how hard they try to look at from another point of view.-MrCrispy (September 19, 2007, 11:41 PM)
--- End quote ---

It's true. But as you know, huge efforts are made to reverse the tendency. Look at PCLinuxOS, Ubuntu, Mandriva, or even Pardus (a very promissing distro from Turkey, funded by the Turkish government). They all have the same goal : everything has to be simple, obvious, easy to configure, compatible, etc., while being much cheaper than windows and while retaining the features that make Linux interesting. I'm still not using Linux on a day to day basis, and I probably won't soon... Unless Vista really messes everything up. That said, I find it extremely healthy for the software world that the Linux community is so alive and growing. "Linux" is truly an amazing phenomenon. And I don't think it will disappear, on the contrary. To repeat what I suggested in another thread: with fast developing countries like India and China (not exactly pro-Microsoft), Linux is probably here to stay fairly popular for a long while.


[0] Message Index

[#] Next page

Go to full version