Exclusive NewsForge report: “The second IBM Linux Scholar Challenge is accepting registrations through October 31. Finished projects are due by December 13. On January 20, 2003, IBM will announce the winners of 20 ThinkPad laptops, and will offer summer internships at IBM’s Linux Technology Center to three of the top winners.“
College Students: Help Improve Linux, Win a ThinkPad
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I follow up on Yamit project, osnews had article on it about two/three months ago.
There was a message on mailing list that they found a sponsor who also donates several laptops (Dell in yamit case) to top technical contributors.
I guess IBM must be pretty cheap after all they are one of the biggest players in computer industry and they make money on linux.
Here’s my !revolutionary! suggestion to help improve linux:
GET RID OF X-WINDOWS, it’s a piece of cr*p.
now…, where can i get that thinkpad.
I would usually be making some witless, random troll here; but I concur completely on a non-trolling basis. X-Windows was alright in the eighties, but to let desktops truly take advantage of hardware accelerated graphics; something impressive for Windowing on the framebuffer must be done, and Cosmoe is one of our few hopes.
Anyone here tried it yet?
I did, a year ago. It was slow beyond belief (I used the SDL output, the rest didn’t work).
But all these are off topic. Back to topic please.
..dont do that. You know you don’t *have* to use X-Windows. You can *choose* to use something else, or not use a GUI. I think many people here are tired of the “get rid of X” argument.
I sure as heck am tired of the “get rid of X” argument.
Have you ever listened to the pro X arguments? Features like network transparency?
Oh! You have no need for network transparency? Wow, i guess that means nobody does. I guess i haven’t been tunneling applications over ssh all this time. It must have been one big X inspired halucination.
I’m also dang tired of the argument that X is pointless just because most people don’t need network transparency. Regardless of whether you need network transparency, or even whether other people need it- it’s a great way to develop a GUI that interoperates between different programming languages and processes securely and effectively.
If you look at other GUIs that are not network transparent, they generally tie you to a specific language (C++ or Java for example) and method of linking your program with the GUI. Network transparent GUIs imply so much extra flexibility that it’s still the best way IMHO to design a GUI.
The whole argument that network transparency inheritly makes a GUI slow is completely bogus. It only means that the commands between client and server pass through a narrow channel like a network socket, instead of a broad channel like a shared library. This enforces the UNIX philosophy of a small number of commands that interoperate well. Just because the client-server commands are passed through this socket doesn’t mean that the GUI will be slow. Generally these commands are relatively low bandwidth. Things that need more speed, like OpenGL graphics or fast bitmap manipulation can use memory mapping. Fast graphics like this wouldn’t make sense over a network anyway, and shared memory provides its own access control methods.
Now, I do agree that X has problems. But they aren’t related to being client-server, and they aren’t related to speed. With proper hardware support X can be just as fast as other GUIs. But it lacks modern features like compositing and proper double-buffering. The future of GUIs probably wont’ be X, but I hope that it’s client-server. Berlin/Fresco still needs a lot of work before it’s useful, but I think they’ve got their philosophy right.
FYI, I’m not a coder for Berlin or X. I’m actually working on my own GUI that I started before I even knew about Fresco. However, it is client-server and has some of the same ideas as Fresco, but designed for greater scalability so it would be practical for embedded systems.
I totally with Micah.
People who dislike X don’t have to run it. (For those OS pros who don’t know, it is possible not to use X)
X protocol is very old and needs to updated. Like Micah said there is a place for built-in support for hardware aceleration.
But, X is great, although outdated. It’s possible to build new GUI that would backwards support X and be very fast.
Since all/most graphic cards support OpenGL/DirectX it’s possible to run GUI in user-space and have it perform great.
On workstation it would simply require giving “X” higher priority so it would be responsive, while no sane person runs X-server on Server either way.
So, I suggest switching “no more X” campaign to something like “no more 3-bottom mice” or “down with donuts”
Cry me a goddamn river, you people can stick with your bloat; but I’ll hold you to your word when something elegant, pretty, lightweight and more responsive than Windows (and, obviously, X Windows) exists built on a framebuffer or direct rendering infrastructure (not implying the DRI project, though it’s possible). I’m sure you won’t so damn righteous about X then
.
I’ll see you in the future, X zealots!
Oh, by the way; Mr. Love might want to work on his grammar, he’s quite illiterate.