Help me help Linux, tell me about Linux problems
Hello, I am a student and I have to pick a project to do for my Bachelor of Science.
I search for problems that Linux has and can be solved programmatically(no community stuff and this kind of things).
No CPU Schedulers... Linux has to much of that.
Proprietary software? Proprietary features? Features Windows has and Linux doesn't?
Any ideas? May be features that don't exist yet but theoretically are possible to implement?
Implementation of KDE technologies in Unity interface. That's what I am going to do myself as well. First, to create "indicators" for Akonadi, Strigi etc. and hook up apps like Amarok or Kmail to existing indicators. Implement features like "cheat sheet" or the fancy window switching effect in KWin. Port Compiz plugins to KWin - no idea how exactly yet - maybe with help of KWin scripting? Integration of KDE semantic desktop inside of Unity. Port of Unity to Plasma interface with QML.
Originally Posted by Alliancemd
How do you like it?
My coordinator won't accept this. It sounds interesting, it reminded me that I was thinking if it is possible to make Unity run on KWin instead of compiz. I consider KWin way more advanced...
Originally Posted by Hirager
Btw, that might be good for you: http://blog.martin-graesslin.com/blo...to-do-awesome/
My coordinator(and me too) wants me to invent something or write an open source alternative to something proprietary which doesn't have an open source alternative(which I find VERY hard, every thing has an open source alternative). Or to write an application to solve hard solvable problems(computational) that will run on clusters...
One guy that has him as coordinator proposed to make an open-source video/audio protocol(a la Skype), the coordinator allowed him to, even if there are some open-source protocols that do that, like XMPP.
I see. I think that enablement of next versions of OpenGL could fit into the "alternative" category. Or anything else of this kind. Also, I read a complaint on this forum about lack of OpenMAX support in Linux.
Originally Posted by Alliancemd
What else... Sometimes I wish OwnCloud sync used GIT as way to resolve file conflicts (not sure if actually solved in recent release - didn't have time to read on that). Also, work on a Wine component could fit here.
How about a proper SIP client? There doesn't seem to exist one that supports all the features at once.
- ZRTP encryption, with integrated easy-to-use verification of the secret string (ie read aloud automatically by a speech engine, waiting for confirmation from the other side - button press)
- encryption for the text messages too
- various codecs, preferably easily extensible should opus etc be good
- light and scalable, ie not in Java, interpreted language of the week, nor using bloated and unnecessary components (be it KDE libs, dbus, ...)
- usable and fast GUI, preferably themeable
- usable from CLI too
- should handle group calls, file transfer
Literally every existing SIP client, be it open source or closed, fails on at least two or more points. We do have many servers that work for almost everything in here.
Maybe helping an existing project that has something missing would be a better idea. ie implement an State tracker for G3D -reaaallly important thing and it might benefit you in the future- or a protocol for one of the IM clients. Or polish a feature in an existing project. I think it will be better than a project that you might abandon after you finish with your studies.
As for one thing thats really missing is a proper CAD/CAE suite but thats not something for a semester project.
That's the problem. We have a like 20 implementations of some things (don't mind the 0 implementations of other things just yet), but they all range from 10-90% complete. They lack polish. It's completely understandable - I'm a programmer myself and I like doing the "interesting work" but I hate putting on the polish. However, as a user, its very frustrating.
Originally Posted by Alliancemd
I think we should take one of the 50 million existing projects and make it better.
(note* all numbers are hyperbole, but they illustrate the point).
Graphics card support for dual cards (specially ati)
Linux lacks graphics support as in windows. Open source drive still fails to load a lot of applications that requires GPU. On the other hand catalyst driver with a tough installation process fails to load any application via wine. So I am stuck on waiting for applications to be natively ported to Linux or dual boot windows for those applications. In older hardware Desktops often freezes (due to compiz). I know windows have moved from xp to 7 and those applications are still slow on windows, but they don't freeze as they do on Linux. I have this problem while running java development environment. I am still on open-jdk-6. While I use maven and eclipse I get almost always freeze when I compile the application. Java's memory keeps on increasing whole day. So I have to shutdown eclipse before I go for launch. I think Desktops should have options to completely turn off compiz (without loggin in to different desktop environment). I even tried to use mint's new desktop environment and used the option to turn of effect, but the solution is still not there. I thought I could load lighter Desktop Environments, but what's the point of using default configuration.
First off, a disclaimer : I'm a hevy KDE user, so most(all) of these will be kde-related.
-Finish porting akregator to the akonadi framework (AFAIK the last ugly duckling in the PIM suite)
-Help the kde-telepathy guys with whatever interests you (contact agregation[my fav.], SIP, UI implementation...)
-Optimize akonadi SQL usage/fix it with SQLite
These are my personal pet peeves, the list of unfinished tasks is quite long.
I'd suggest you take a look through GSoC projects that weren't accepted, and take your pick from there.
Other than that an area that would interest everybody are graphics... I'm sure te mesa guys would find you something
I have to agree with 89c51 that creating a new project just for the sake of it is IMO a bad idea (well, you will certainly learn, but other than that it will be useless). You want to help the OSS landscape, join an existing project (unless you really find a part of the landscape missing).
This! There were a few attempts here and there, but nothing usable ever materialised. And the proprietary ones usually don't support Linux as well as cost way into thousands of dollars.
Originally Posted by 89c51
Blender showed that it's possible to make a good alternative to proprietary 3D modelling programs, so a proper CAD tool should also be possible to create.
Tags for this Thread