Slashdot Mirror
Cygwin/XFree86 Leaving XFree86.org
An anonymous reader writes "The Cygwin/XFree86 project is leaving XFree86.org. For those that don't know, Cygwin/XFree86 is a port of the X Window System to Cygwin (which provides a *nix-like API on Windows). Here is the announcement and the start of the trouble. The XFree86 project has pushed away more developers than most projects ever have - is this the beginning of the end for XFree86?"
These guys seem to care more about being able to brag about their commit access in their email signatures than streamlining development of their software and making things as easy as possible for those willing to devote their time and talent to the project.
If ever a project was in need of a fork, and if ever some project developers were in need of an attitude readjustment - this is it.
And he presented his case well. These other XFree86 guys sound like the cast from Othello... way too serious for what is, after all, something that's supposed to be fun: working on an open source project.
You know, what kind of nut must it be to crack to get X running atop of Windows? You'd think they'd give Harry some slack just out of the complexity of what he's doing.
Another poster mentioned that it's the fault of the tools, and I think this is a good point. A truly usable code management system would allow for Bozo the Clown to have commit privileges and it wouldn't impact the overall effort at all.
Is this truly the only Earth I can live on?
I think it would be best if all these projects that left Xfree86.org united - the Cygwin/Xfree86 folks, Keith Packard, and pull their resources to come up with a workable development model (yeah, and a workable X - all major projects - Gnome, KDE - are waiting for long promised features that all modern graphical subsystems exhibit except for X.)
Actually it was less to do with "Bloating it" and more to do with fixing bugs.
But then I read it.
CJC
The XFree86 project has pushed away more developers than most projects ever have - is this the beginning of the end for XFree86?
Why the scaremongering, anonymous submitter? Just because one project isn't getting access to XF86's CVS tree and will have to maintain one of their own somewhere else, doesn't mean that everyone will abandon XF86. It's mature, has a ton of features, and has no viable replacement; who is gonna leave and where are they gonna go?
the coolest club on
From their Savannah website:
Xouvert is a development branch of the Xfree86 source tree. It's purpose is to provide wide testing and integration for third party patches, and to test and stabilize innovative new ideas for submission to the main Xfree86 branch.
Doesn't really sound like Xouvert marks the end of XFree86. Indeed, it sounds like Xouvert is dedicated to improving XFree86.
Dinivin
I wonder... do people say "X is going away" because they think it is, or becasue they hope it is?
#define DRM chmod 000
Remember when Gnome split from KDE? They fully intended to end KDE, yet today both are powerful desktop systems that have benifited from each other. (Last I cheked you can't even complile KDE with a couple GNOME libs - code reuse in action)
For that matter, linux was the end of BSD, or perhaps we should say OpenBSD was the end of NetBSD. Take your pick of history, BSD is alive in well despite what some anonymous cowards would have you believe.
This is a good thing, XFree86 has gotten a lot of criticism, let the critics go their own speerate ways and each prove their way is best. In the end the best way wins, or if there is no best way, all survive, and each focuses on the areas where its way of doing things is best.
What advanced features don't you need?
Most likely, modern desktop users need much more of the advanced features (recent extensions such as video, OpenGL etc., in particular) than ever.
If you look at what the basic X11 feature set really is, it's really very simple.
Most likely the most complicated thing you aren't using is the color management stuff.
What most people experience as "X11 bloat" currently probably consists more of bloat on the widget toolkit side than on the XFree86 side.
XFree86 could use a lot of cleaning up, but it's not particularly bloated.
The reason the push developers away is that many of these guys are trying to bloat xfree to hell
Have any proof to back that statement up?
Harold was requesting CVS commit access only for bugs that pertained to Cygwin only -- they had no impact on other platforms. Hell, if properly ifdef'd they wouldn't even compile into the binaries on other platforms. That doesn't mean they're not bugs though, and it doesn't mean they shouldn't be fixed in the main tree.
We're not talking about features here. And there's a long line of people that have tried to get XFree86 to fix bugs -- either in the core or in drivers -- that have not only been denied commit access but also had their fixes ignored, their questions ignored, and been passively shoved aside when trying to get things fixed. The number of actually active developers (i.e. - number of people with commit access and are actually spending time on the project) on XFree86 is absurdly low for the size of the project.
I would be willing to wager that >75% of those of us who run a Linux desktop don't need hardly *any* of the advanced features in the X Windows server.
I would wager that >75% of all software users don't need *any* of the advanced features of the software they use on a daily basis. I would also wager that those in the 25% range drive over 95% of the innovation and development, and that those users _need_ (as much as anyone can need anything) those advanced features.
Doesn't anyone know about the 10/90 or the 20/80 rule anymore? (If no, look it up).
I would like to see a completely modular, X-windows core-compatible windowing system for Linux. Want to use some of the advanced features? Add in the module, recompile, and go!
1st, to me, modular means you don't need to recompile. 2nd, who really cares how modular X is? That surely wouldn't help me get cut and paste working (by this I mean between all X apps and beyond text data). That surely wouldn't help me get drag and drop working. These little features that are over 20 years old are welcomed!
Until these basic needs are met, I don't want to hear another "Is Linux ready for the desktop?" questions.
Xouvert is a development branch of the Xfree86 source tree. It's purpose is to provide wide testing and integration for third party patches, and to test and stabilize innovative new ideas for submission to the main Xfree86 branch.
It's an interesting phenomenon associated with free software: enough talented developers get the perception that the current people in control are being unreasonable about release schedules, overall direction, features, bugs, indenting styles, etc. and fork their own branch.
A closely-related parallel here is how the egcs folks wanted greater ability to change the gcc codebase than the gcc developers wanted to do.
Then, the egcs branch took off so famously that later it essentially became the gcc development branch.
May the best X branch become the tree trunk.
"Provided by the management for your protection."
If you actually followed X or even the discussion linked to for 5 minutes you'd realized that it is NOT the problem at all.
The problem as cited through the list is that the core developers do not allow external commits by major commiters like the entire Xwin cygwin port.
These people have to wait weeks for any changes to possibly show up in the CVS because the core developers don't have time for it.
The core dev's answer: Shut up and stop complaining we are doing the best we can.
This has nothing to do with bloat and everything to do with control and workload.
Bye!
I certainly can't speak for XFree86, but this is normal practice within the engineering group at my employer, and AFAIK at other commercial development houses. Raising a bug for a feature means that it can be tracked as a commit, means people can make comments on it, means you have a common format for all commits, be they bugfixes or new features, and so on. No, I don't really like it either, but it makes a lot of sense.
Paranoia isn't an infectious condition, it's a way of life
The X network layer is not a "bloated" bolt-on kit or added feature that someone wedged into X as a gimmick... X is itself just a specification for a data stream, like umpteen other protocols you have in your /etc/services file. At its core it is really quite simple and lightweight.
Furthermore, when the client and server are on the same machine, the data stream is NOT sent over the network, but is routed through local UNIX sockets or shared memory, making X essentially as slow or as fast as your system bus and graphics hardware. Only when you actually separate client and server on to different machines does X use the network sockets.
Overhead is simply not a factor on an average Linux desktop.
This feature bloat everyone is frightened of is in other places, like for example the KDE and GNOME architectures and the desire of most users to drown in pixmaps and theme engines.
With that said, on my own Linux desktop (a lowly 900MHz PIII) I use KDE 3 and play Quake III and so on and I don't find it to be any slower than Windows 2000.
Maybe there is just a small crowd (the ones who keep submitting "3D site" or "hardware site" stories) who won't feel elite at LAN parties until their Linux box can beat Windows boxen by at least 6fps in frame rate tests, 403fps. vs. 397fps.... and they're somehow convinced that if they can just get rid of that damn protocol and somehow drop "abstract" graphic ideas directly into video memory rather than organizing and processing them, that extra 6fps will be forthcoming.
Meanwhile, the rest of us continue to use the god-send network features of X to administer large installations from a single point of access, or to deploy narrow-application thin clients at greatly reduced cost.
STOP . AMERICA . NOW
The context to harold hunt saying "go fuck yourself" to Thomas Dickey:
Thomas Dickey says:
"well, when you graduate and (presumably) find a real job, you'll have a chance to get an idea of where time goes. the patches _are_ applied, right?"
Which is an extremely rude thing to say to anyone. Even more so when Harold has already spent several emails explaining, and also is apparently currently suffering some fairly serious medical problem.
"When you are in a graduate degree program and working 30-40 hours per week, that is a *lot* of time."
"Seriously, I don't know why I waste my time submitting patches that are specific to my platform and then wait up to three weeks for them to be committed."
"Can I please finally be given CVS commit access with the understanding that I am a moron and that I will only commit things that are cygwin specific,..."
All he wants is the ability to commit to CVS for the module that he is the expert on, and David Dawes and Thomas Dickey are unfriendly, insulting and rude to him. Not exactly a good way to run an open-source project. Did they not read the Cathedral and the Bazaar?