How To Help & Support Linux, Open-Source?
Phoronix: How To Help & Support Linux, Open-Source?
If you've been wanting to get involved in supporting Linux and other open-source projects with or without a programming background, or you have creative ideas how to get involved, see this thread...
The kernel could indeed use some work:
There are many things that aren't done in a unified way, thus defeating the purpose of it a little to completely.
e.g. API and kernel code for things like clipboard, input keyboard and mouse with infrastructure for keymaps and other stuff.
And of course (this is just my taste) the way how the directory structure is done is horrible.
All distros should take a look at how GoboLinux does things.
I disagree. Gobolinux does things with help of myriad of symlinks, beside this such tecnique is completely unneeded. FHS already does fine job by spreading the app in the system and package manager goes on the rest to track the spread should it be upgraded or removed.
Originally Posted by plonoma
But this is much nicer way: http://brainstorm.ubuntu.com/idea/6243/
And I think it is filesystem job - we could have filesystem assign each file either to directory FHS structure or similar to those in 6243; and at same time mark each file as belonging to specific package - say owner "package", additional attribute "packagename-version".
Providing easy access to user-facing features is the job of the userspace. Why should this fall to the distros implementing ugly filesystem hacks that break pretty much every script ever written in the 33-year history of the things?
How do users access applications? From the GUI application list (most often DE-provided and updated; dark magick), from icons on the desktop (I think people may still do this) or in various docklets (added manually, often by dragging and dropping from aforementioned menu), or from the shell (also has powerful magicks, like $PATH).
Knowing the directory where the application lives isn't important. It's superfluous information for the user; useless.
It's like all the people talking about "Stable ABI! Stable API!". The kernel syscall interface is stable for userspace applications. What does stabilising those things improve? Judging by the nightmare of partially-working legacy interfaces and cruft, not much. To me, it says they've never heard of Windows viruses or upgrades that break hardware compatibility. I can't imagine any of them would like that same situation for Linux, and yet....
If you want to help Linux (in a concrete sense), but aren't a programmer, we really do need artists and designers (especially UI). Or even just people with enough brains to write more than "It broke when I clicked it".