Fedora Prepares For Xorg Instead of XFree86

ZuperDee writes "I noticed in the development branch of Fedora today that they appear to be in the process of creating new xorg RPMs, and from the looks of the changelogs in those RPMs, it looks like their ultimate plan is to switch from XFree86 to the XOrg Foundation's implementation of X11. Anyone else here think this could signal the beginning of a new trend in Linux distributions, and that XOrg could end up becoming the new de-facto X11 implementation?" (See this earlier story,too.)

RE: Drivers

[Professor Cool Linux]

this is a fork so it should be compatible

Re:drivers

[Lussarn]

The X.org server is basically a branch from Xfree just before the licence change. They should be very similiar at this point.

Path of least resistance

[nonmaskable]

I think the XOrg codebase is pretty much the last pre-license-change (4.4rc2) release, plus work done by the folks recently run out of XFree86 by Dawes.

Re:For the ignorant (like me)

[sethadam1]

Xorg is an implementation of X11. XFree86 is another implementation of X11.

What you want is to know the difference between XFree86 and Xorg.

Re:First X Sucks Post!

[Tet]

James Gosling, X Architect

Of course, Gosling was never an X architect. Those were Scheifler, Gettys and Newman. Gosling was the architect of NeWS, a competing windowing system that ultimately lost out to X. Yes, IHBT. Thank you and good night.

Re:For the ignorant (like me)

[DA-MAN]

Could somone go over the diffrences between X11 and Xorg? Is it just a license issue, or are there other differences?

X11 is the 11th iteration of the X protocol. XOrg, XFree86, and most commercial X servers speak X11R6 these days. Speaking the X protocol is key to interoperability from Unix to Unix.

X11 as a protocol doesn't have a license issue that i am aware of. Did you by any chance mean the differences between XFree86 and XOrg?

If that is what you meant, then the answer is simple, XOrg is a branch right before the XFree86 license change, so it's pretty safe to say that XOrg isn't too different at all at this point in time.

Re:First X Sucks Post!

[FooBarWidget]

That's what the XRender and Xft are for! They are full replacements for the old rendering model and font subsystem.

Re:This is great news.

[AKnightCowboy]

I am glad to see them choose X.org over freedesktop.org. I do not want to see X be replaced with an LGPL fork.

Hmm, the link is to xorg.freedesktop.org. Are you sure you got what you wanted? It looks like they ARE using the fork.

Re:For the ignorant (like me)

[Omnifarious]

To expand more helpfully on the previous poster's comment...

XFree86 and XOrg are both implementations of X11. X11 is technically a protocol, not a particular program. This is why X11 has persisted for so long despite repeated attempts to dislodge it. Everybody who tries to do something better forgets that X11 is a protocol, and that's actually why it's so popular. They usually end up implementing something that's an API, which is just all wrong.

The XOrg implementation of X11 is a fork of the XFree86 codebase, just before XFree86 changed its license to be not quite free enough for most people to be comfortable using it.

Re:If it's compatible, they will use it ...

[vidarh]

Considering that it's the X.Org foundation that is maintaining the X11 standard, the compatibility is a given - their X11 implementation IS the reference implementation of X11.

Proper Context

[rjstanford]

The quote, as seen in the actual article:

"They were going as far away from my design as they possibly could," he said.

How so?

"The color model and the fonts is aggressively stupid. It works, but oh my god. It's awful."

Here its obvious that Gosling not only didn't create the color model and font system that's part of X, but in fact was proposing quite different solutions as part of his NeWS competitive system. The parent quote makes it sound as if he's admitting that he created them in X, and now hates them.

C'mon now...

Besides, if you never read the articles, and just look at the exceprts, you'd never know about the asparagus. What asparagus, I hear you ask? My point exactly.

Re:you might want to check this out

[Lussarn]

There are two Xservers at freedesktop.org, the one this FAQ goes to is not the one implemented in Fedora core. The one in Fedora core is a fork of XFree. The one this FAQ is for is a newer and interesting one albeit not ready for prime time yet.

Re:De Facto

[DrSkwid]

lol0x

ever have an MSDN subscription?

finished? you gotta be joking

documentation? that's whats source code is for

Re:Great

[FooBarWidget]

Who modded up this as "Insightful"? It's nothing more than clueless bashing.

X11 is a standard, not an implementation! Just like HTML is a standard!
That distro A uses XFree86 and distro B uses XOrg means absolutely nothing to end users. Everything's still interoperable because X11 is a standard. Everything will still Just Work(tm) and the end user won't even notice something has changed.

Re:Drivers could be a problem for a long time.

[tannhaus]

I can't speak for ATI, but I know that nvidia includes the source for their drivers. You can actually compile it on your machine. So, I don't see this as really a problem with nvidia...especially since X.org code is just a fork of XFree86 code. Just recompile and use.

Re:For the ignorant (like me)

[Crispy Critters]

"Speaking the X protocol is key to interoperability from Unix to Unix."

How about "key to interoperability between X client and X server". Remember that X was implemented on VMS as well as on Unix, not to mention the version in X terminals and various emulators for MSWindows and Mac.

Re:Drivers could be a problem for a long time.

[Nerull]

Thats the source for the kernel interface, not the X driver itself..thats a binary.

Re:Drivers could be a problem for a long time.

[EnormousTooth]

Umm... as far as i know, you CAN use the existing drivers with it. It's a fork of XFree86.

Re:Drivers could be a problem for a long time.

[Senjutsu]

Xorg is a barely couple-month old fork of XFree86 that took place right before XFree switched to their new lisence. Driver compatibility isn't going to be an issue.

Re:Drivers could be a problem for a long time.

[SillyNickName4me]

Look again.. they provide an OS wrapper in source form, and there is some odd binary file in there that has the actual driver.

OK listen up

[Anonymous Coward]

Alright, there's been a shitload of ignorant posts here.

First off, this new server is a snapshot of XFree86 just prior to the licence change. Basically a fork.

Second, it basically has nothing to do with X.org - I don't know why they call it that, most likely due to the licence.

Third, X11 is the protocol that X servers speak nowadays. X version 11 release 6.6 to be more precise.

Fourth, nvidia and ati drivers will work.

I hope this clears it up somewhat.

Re:I'm confused

[Fourier]

Xouvert differs from the others in that it appears to be a dead project [gnu.org].

Re:drivers

[metallikop]

Yes, they're very similar but X.org is mainly a testbed for ideas. That's fine and all, but X.org doesn't even have a stable branch. I've tried it out and it runs a bit slow right now, and a refhresh rate of 60hz is kinda hard on the eyes. Last I checked that's what you were stuck with.

Re:This is great news.

[Rick and Roll]

Actually, both projects are on Freedesktop.org. One's called Xorg, the other Xserver. And Xorg appears to be under the standard X license.

Kind of strange, but not really. Just one project (freedesktop.org) providing excellent free CVS hosting for free desktop projects, and two very similar projects with very different leadership joining.

The Lowdown

[Anonymous Coward]

Ok kids, here's the quick summary to get everyone up to speed.

XFree86 [xfree86.org] and FreeDesktop.org's X server [freedesktop.org] are both X11R6-compatible X servers. The FreeDesktop.org server (herein known as XOrg) is a fork of an old XFree86 project called the KDrive.

The KDrive was a tiny X server implementation originally designed for PDAs and such. When you compile it the binary comes out to about 700kB and it requires hardly anything else to function. The author of the KDrive took (read: forked) it from XFree86's tree and started adding onto it, and it became XOrg.

So X11R6 applications and libraries work almost exactly the same under XOrg. The XFree86-specific extensions to drivers and shit need to be ported but most apps don't use those.

Gentoo, RedHat, I think SuSE and Debian and soon to be more Linux distros are all slowly switching to XOrg. Until then they'll be shipping XFree86 4.3.99.902 and below as those are the ones without the evil licensing changes.

This has been in the works for some time people, so it's not a rumor or a guess.

Note: XOrg isn't the real name of the server, I just call it that cuz im lazy. XOrg is the name of a foundation that puts out this FD.O Xserver. Info here [freedesktop.org].

Re:This is great news.

[be-fan]

Let's stamp out this rumor before it spreads further. The new FD.O X server is under the standard MIT X license, not the LGPL!

Re:drivers

[be-fan]

Methinks that you're confused. X.org is seperate from the FD.O X server. X.org is a derivative of XFree86. FD.O is different.

Corrections

[Alan Cox]

Umm ok

First its based on XFree86 4.4 just before the change, with the non-contaminated further changes added and other stuff not in XFree 4.4

Secondly it has a _lot_ to do with X.org. The wheel has turned full cicle from when years back OpenGroup/X.org tried to change the license and XFree basically told them to go away to today where X.org is doing the same thing the other way around and keeping it free. X.orgi is part of this now.

NVidia and ATI drivers may work. The Nvidia ones at least are reported ot do so, although they have chronic problems working with the preferred kernel build settings like 4K stacks.

Re:I'm confused

[a.ameri]

X.Org is something very new. It is just a complete fork of Xfree86 4.4 rc2, which was the last version before the license change. X.Org's sole reason of existence, is the license change in XFree86.

Xouvert is also based on Xfree86, but it is a bit different than X.Org. Xouvert was started when it became apparent that XFree86 guys were too reluctant to change, and to commit new codes and technologies. If I am not mistaken, the Xouvert project started in summer of last year, with the goal of being a more experimental branch of Xfree86 i.e: they would accept code more easily than XFree86 guys. They also stated that they want to seperate the drivers from other parts, so that one can add a driver of a new chip, to a old release of X. I don't know how succesful they have been in this front.

And aside from all of these, is the Free Desktop.org's X Server. This X Server, mostly written by Keith Packard is not mature for every-day use yet, but I think of it as the future of Open Source X. It is mostly a complete rewrite, and it is not a fork of XFree86, though it has borrowed some libraries from the latter one.

Re:For the ignorant (like me)

[pe1rxq]

That is why you run it over ssh if you have to cross an unsafe network...

Jeroen

Re:drivers

[floamy]

You are way wrong. The freedesktop.org xserver is based off kdrive. They have already implemented alpha-transparency and stuff via their new xcomposite extension.

Re:I'm confused

[be-fan]

Let's seperate the code-bases from the organizations. There are a couple of organizations:

XFree86, Inc. - The old organization, mainly consisting of David Dawes at this point.

Xouvert - Splinter group that forked X awhile ago, with the intention of being a cooperative competitor.

X.org - Formerly X Consortium. Bunch of companies and developers working on the X11R6.x reference codebase.

Freedesktop.org - Umbrella project for various desktop-related Linux projects

Now, there are some implementations:

XFree86 - De-facto standard on Linux, by XFree86, Inc. Based on the X11R6.x reference codebase.

Xouvert - Fork of XFree86 (circa 4.3?) by the Xouvert project.

X.org server - Don't confuse this with the X.org reference codebase. This is a fork of XFree86 4.4-RC2 (before the license change). Now its under the X.org umbrella, and is hosted on freedesktop.org (that's the confusing part :)

FD.O X - Keith Packard and friend's new, fancy X server. Development hotbed for new technologies like transparency, OpenGL-acceleration, etc.

There are a couple of seperate sub-components to note here. The FD.O X server supports a number of DDXs (basically, driver layers). There is the kdrive-based DDX, the XFree86-based DDX (called Xizzle, theoretically compatible with XFree86 drivers).

There will eventually be another DDX designed from the ground-up for OpenGL acceleration. The device-independent portion of the FD.O server is, IIRC, derived from an older version of XFree86.

Re:If it's compatible, they will use it ...

[be-fan]

Except this isn't the reference implementation, its XFree86 4.4 pre-licensing change. However, since XFree86 is based on the X11R6.6 reference implementation, it should be very compatible.

Re:drivers

[cortana]

I believe you are thinking of Keith Packard's X Server. Freedesktop.org has another X server project, forked from xfree86.org's 4.4rc2 release.

I think.

Re:drivers

[be-fan]

Didn't I say as much?

The X.org X server is XFree84 4.4-RC2 + bits. It is hosted on freedesktop.org

The FD.O X server is KDrive (which is derived from XFree86, in part) + bits. It is also hosted on freedesktop.org.

Re:drivers

[be-fan]

Cuz its confusing. What you ran was not the X.org X server, but the freedesktop.org X server. Its where all the fancy transparency stuff is being developed. To support that stuff, they'll have a new driver model designed to take full-advantage of OpenGL. Since those aren't ready yet, they're using the kdrive model, which is where the low-refresh rates and general lack of acceleration come in.

The X.org X server is the XFree86 4.4 codebase, so it is binary-compatible with the ATI/NVIDIA drivers.

Re:The Lowdown

[IamTheRealMike]

Sorry, that's totally wrong.

There are two X servers at freedesktop.org now, both with stupid and confusing names but hey :)

1) Xserver - this is the new experimental one that does pretty drop shadows and stuff. Not really mainstream yet. This is the fork of kdrive.

2) Xorg - this is the fork of XFree before the licensing change. It's not experimental and is usable just like XFree is.

Hope that helps

Re:I'm confused

[be-fan]

There are two parts to the FD.O X server. The DIX (device-independent) layer is derived from XFree86. The DDX (device-dependent) layer is new. The libraries are derived from XFree86 too.

Re: Drivers

[dmoore]

It was about more than the license change.

The main problem was that many folks got fed up with the very closed nature of XFree86 development. Many decisions about the project were made by fiat in non-public mailing lists. These core group of developers were often unwilling to explore new features or allow new developers. The barrier to entry for obtaining CVS access to the source was high. Thus, many developers who were not part of the core group got annoyed and decided to stop submitting patches to XFree86. Thus, all these derivatives were born that promise a more community-oriented development process.

The license change was just the straw that broke the camel's back.

Re:De Facto

[mabinogi]

The GPL does not prevent forks, and no one at the FSF will claim it does. In fact a large number will probably tell you that one of the benefits is that it allows Free Software forks.

What it does do, is prevent non Free forks.

Troll, mod down

[Anonymous Coward]

This is clearly a troll. None of the code in question is covered by the GPL or even LGPL. Besides, the move shows a clear trend towards developers moving in the more-free direction, not in the less-free direction.

Re:Core 2

[randomblast]

No, FC2 is already in beta, so they are at the stage of getting rid of bugs.
New features will be saved for FC3.

Re:De Facto Standards

[irc.goatse.cx troll]

" I usually need both installed if I want to be able to run any application without too many problems"

I've neather installed, and everything I run runs fine. All you need is the qt and gtk libs and you're fine.

Re:Ad-hoc Standard

[Etcetera]

On other distribututions, rpm (yes, rpm) works fine. it doesn't strip binaries it shouldn't touch (arm for instance in an x86 package), it doesn't add depenedencies it doesn't need, it basically just works as advertised.

Have you used the 'AutoReqProv: no' line in RPM? Works fine for me in preventing spurious dependancies.

there is no such thing as a minimal redhat9 installation. all we wanted was to build packages for redhat9, 2gig was as small as I could get the build.

Have we not heard of "Select individual packages?" I routinely build RH9 and FC1 boxes in the 800-900MB range. Could probably do with less if I really wanted to.

There are also the configs, the different command usages, etc. etc.

When you're using any package management system there are bound to be configs that are placed in automatically. Aside from RH basically not using /usr/local (something that I agree with), I don't see what you're complaining about.

Redhat: new users and coporate users needing a good backup plan.
etc...

Blah blah blah. It's certainly as possible to tune a RH system for low-fussness, or for high performance (i386 packages don't make that much of a difference!) just as it is to mis-configure any of the other distros. Use what works for you.

Probably not.

[Andy Dodd]

I'm expecting that the majority of distros will very quickly follow Fedora.

I know for a fact that Debian, Gentoo, and a few others are specifically NOT touching XFree86 4.4 (i.e. post-license-change), and are looking for alternatives.

X.org sounds like it is currently the most mature alternative, and will likely have the marketshare XFree86 does within months, unless David Dawes pulls his head out of his ass and stops shooting himself in the foot. He doesn't seem to realize that his license change is going to make XFree86 a defunct project VERY quickly.

Re:Y-Windows

[SETIGuy]

Give me one good reason it's not a stupid waste of time and effort and I'll change my mind.

While the freedesktop.org screenshots are pretty, they ignore that X11 was developed too long ago. To many of those items in the pretty pictures would, on most X servers, give messages ranging from "extension 'this_weeks_version_of_something_like_render' not found" to "SIGSEGV".

X has some serious problems. Too much functionality has been put into optional extensions. Not to mention that widgets and toolkits should be part of the server, not compiled into the client. We've learned some things about windowing system design in the last 25 years.

The freedesktop demos wouldn't look so good if the server was running on Solaris and was displaying clients running on the other side of a 56 kbps link.

X has outlived the usefulness of its design. It's time to move on.

Re:Drivers could be a problem for a long time.

[Trejkaz]

And besides, there is a driver layer for the freedesktop.org Xserver which can supposedly use the XFree86 drivers. In theory, at least. Once it's stable. ;-)

xorg-x11 works great

[dpw2atox]

I currently installed the rpms by Mike Harris from redhat/fedora on my Fedora Core 2 Test install and it works great. I simply had to reinstall my nvidia drivers, which work fine for everyone out there worried they won't work, and it runs fine. I for one am glad they are making this transition. It is time that X be open and maintained by a community with bugzilla. As more and more patches are sent and applied more companies may produce patches for their hardware since they are actually being accepted. This in my opinion is going to do nothing but help and improve a users experience with linux.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Comment removed

[account_deleted]

Comment removed based on user account deletion

Not all about licenses

[0x0d0a]

It would be a nice demonstration of the claim that opensource software can adapt quickly to 'breaks' in incompatable licenses (and unwanted behaviour).

This is not all (directly) about licenses. Keith Packard has done most of the new, interesting functionality in XFree86 for some time. By going with him, they are aiming for more modern functionality in their X server. XFree86 is very conservative about new functionality.

Re:For the ignorant (like me) origin of X11 name

[Piquan]

X11 was X11 right from the start as far as I remember. The 11 stands for one megapixel (as in a display 1000x1000) and one MIP (million instructions per second).

Sorry, the 11 is a version. From "man X":

The X Consortium requests that the following names be used when referring to this software:

• X
• X Window System
• X Version 11
• X Window System, Version 11
• X11

I'm guessing your megapixel*MIPS was a retcon. Some of us are actually old enough to barely remember when X10 was just passing out of relevance, and I'd imagine a few of us remember before that. Versions before X10 were never really relevant outside of MIT. X10 was 1986, X11 was 1987, and there's been various X11R*s since then. Today, we use X11R6.4, but many programs want lots of extensions on top of it (eg, XRender). Since many of these have only been implemented on XFree86, that's now a de-facto standard.

Re:Questions from an illiterate X user

[0x0d0a]

X.org is the more conventional fork (and Xserver is the fancy one with transparency, drop shadows, etc), both at freedesktop.org.

Keith Packard is, I believe, the freedesktop.org guy -- he's the guy that's pushed forward most of the new XFree86 functionality for the last few years, until his falling out with Dawes.

This is an annoyance for packagers. It will also mean that transitions to Xserver or features coming across from Xserver are more likely, which is good all around for Linux folks.

Re:drivers

[paule9984673]

Actually, there are at least three:

• Xorg (Xfree86 fork by X.org)
• The one based on Kdrive
• Xizzle (Xfree86 fork by some other people)

Re:Y-Windows

[d3vi1]

Why abandon one of the greatest technologies ever created in computer world?
X-Windows is, just as it says, a server.
One of the greatest things about it is it's network transparency. X-Windows, is still ahead of it's time. Microsoft introduced Terminal Services back in 97 or 98 for Windows NT Server TSE, long time after X-Windows existed, and it still is not as powerfull as X11, it only draws the whole screen through a pipe, compresses it and sends-it to a client. X11 does a lot more than that, it has security is a number of forms (e.g. ACL based), it has support for extensions - which is soo great, and it tells the client which extensions it supports, it has speed (when not over the network) using UNIX Sockets instead of TCP. Even over the network it's fast. If you think that running mozilla remotely on a 56k is slow, think of the alternatives.
Also XFS is great. Imagine you're in a DTP office. You need hundreds of fonts, an UNICODE font can have 20MBytes, or more, why should those fonts be copied on all the stations? One central station for all of them is enough.
You want remote desktop? Just thing XDMCP.

X11 should NOT have an integrated widget set in it. That is because, it's multi-os, multi-platform, you can't expect all the platforms to have the same widget set, toolkit, just think embeded devices here. Not to mention that there is already a standard widget set as defined by IEEE(or was it ISO?) standards: motif. Unfortuantely motif is getting kinda old.

Re:For the ignorant (like me)

[GuyWithLag]

Ermm... You *did* know that on Linux and most *BSD's a context switch between two processes is faster than a context switch between two threads (of the same process) in Windows.

The real reason that X11 is seen as slow or laggy is that window resizing/moving leaves a trail of damaged window contents that are momentarily visible to the user. Solving this problem is non-trivial, as it requires either global transactional-like synchronization or aggresive buffering, like the recent reseach into translucent windows has required.

Re:drivers

[bfree]

The following is shamelessly lifted from Daniel Stone's blog which is aggregated on planet.freedesktop.org [freedesktop.org].

Daniel Stone: the various x trees: an explanation

OK, listen up kids, 'cause I'm only going to say this once. I think.

DIX: Driver-Indepdent X. Anything that isn't server-specific (extensions, core functions/structures, et al), goes in here.
DDX: Device-Dependent X; the inverse of the above.
XFree86: They still exist. They just released 4.4. The XFree86 distribution contains a DIX, a DDX, X libraries, X apps, fonts, and docs.
X.Org: They have a tree forked from XFree86 that contains all the same stuff, and still uses imake. They're working on a release. They're open and stuff now.
xserver: The freedesktop.org xserver project has a DIX and three DDXes - KDrive, Xizzle, and XWin. That's it.
KDrive: 'Keith's Driver', formerly known as TinyX. A completely separate DDX to XFree86 - very small, used as a testbed for stuff like RandR, Composite, Damage and Fixes. Good for embedded machines.
Xizzle: A fork of the XFree86 DDX, built with autotools, et al. Only just starting to link an actual binary, doesn't work yet, but is moving very, very quickly. Also, the binary is half the size of XFree86's. Pronounced 'shizzle'; mea culpa.
XWin: Cygwin's server for Windows.
freedesktop.org: We have xserver for the server, xlibs for the libs, and xapps for the applications. Everything's modular: the release schedule of the ATI driver is no longer tied to that of the X wire protocol, or some random fonts. Word.

What's interesting to note is that Daniel Stone is the person who did a lot of the work on XFree86 4.3 for Debian and became co-developer of XFree86 for Debian. He is now very active on fdo (seemingly focusing on Xizzle) and also Keith Packard is becoming a Debian Developer. So if Fedora is looking like it's going for X.Org, it looks like Debian might be going to fdo! Truly I think everyone will remain on a forked XFree86 (possibly X.Org) until fdo is "ready", the question being what will the binary driver developers do?

Answers, from the paper on the site

[bonch]

Why abandon one of the greatest technologies ever created in computer world?

It's not one of the "greatest technologies ever created in computer world." You've got to be kidding me. Then you go into a long advertising spiel on X11.

Anyway, here are the reasons listed in Mark's paper:

"The X Window System [23] is the de facto standard graphical user interface (GUI) system on UNIX and UNIX-like platforms such as GNU/Linux. However, as X approaches its 20th year, signs of its age are beginning to show. Commonly cited problems with X include:

• X is too slow. This is commonly dismissed as nonsense due to high throughput that tweaked implementations of X have been proven to achieve. What this does not take into account is that in the general case it is latency that matters more than throughput [6]. Unfortunately, the design of X does not facilitate low latency.
  
  
• X places too much burden on the programmer. The X protocol, and its corresponding library Xlib, provide very low level operations. As a result, programming directly with Xlib is very difficult. For this reason, programmers usually choose to use a toolkit library.
  
  
• X has no standard toolkit. In 1984, before GUIs were common-place, not providing a standard toolkit was the best way to achieve enough flexibility to create all the applications that had not yet been conceived. However, these days, with the benefit of the last two decades of experience [16, 25], it is much better ot provide a complet eset of standard user interface components that look and behave consistently.
  
  Aside from the user interface inconsistency, the lack of standard components also makes internationalisation difficult, particularly for languages which require a complex input method.
  
  
• X is reaching the end of its life span. XFree86, the most popular version of X that is in use, is now over 10 years old. Over the years it has been extended and modified many times, to the point where it is an incoherent mess.
  
  Although the X protocol supports extensions very well, some of the latest extensions have begun to interfere with each other. For example, when Xinerama (the extension which allows X desktops to span multiple monitors) was first released, it broke XVideo (the extension which allows X to use hardware accelerated overlays for video play back). The 'fix' for this was to allow XVideo to only work on the primary display. The latest extension, XRandR (Rotate and Resize), is also known to break many older applications which assume that the screen size will never change.
  
  Further, the internal design of X itself is outdated. Even adding a simple feature, such a stranslucent windows, requires large changes to the server [17]. Because of the requirement to be backwardly compatible, these features must be implemented for everything that X works on, including two-colour displays.
  
  
• X is too complex. The years of extension and modification of the X protocol itself hav eleft he unfortunate legacy that X is too complex. Additional protocols like ICCCM which have been layered above X in an attempt to solve problems have caused additional problems when it comes to understanding what is actually happening [24]. The xine media player for Linux has to probe which window manager is currently running and guess at the best way to switch to full screen. The developers gave up trying to find a consistent way to switch off the screen saver, and switch to the ugly hack of simulating ht eleft shift key being pressed once every thirty seconds [7]."

Re:Answers, from the paper on the site

[d3vi1]

1) X is too slow. It's not a problem of X11. It's a problem for the toolkits that use X11. If the connection is a 56k, why the bloody hell, doesn't the toolkit give-up some of that eye-candy. Well here it's the user's fault also. X11 does not require you to use 24bpp. XFree86 actualy accepts 1,4,8,16,24 bit images. Think mozilla looks slow over 56k? Try netscape navigator then. It's not as slow. No matter the protocol you transmit the data remotely, it's gonna be slow over 56k.

2) No one said that you should use Xlib. I personally think that the ones who succed in that, are both heros and masochists. Ever tried using a toolkit? Say GTK or QT or TCL/TK or motiff...etc? Most of these have nice language bindings. For GTK, you have also the very fast C and the soo easy to use C#.

3) There should be no standard toolkit. You cannot use GTK or QT on a palm or something like that, first because they are big, and consume a lot of that precious space, second because all that eye-candy doesn't fit on the display, third because they consume too much CPU power. The thing with the input methods, it's very well defined and integrated in X11, xinput works perfectly. Localization and accesibility should belong to the toolkit because they are soo related to the way the program is written.

4) X11 is a protocol not an implementation. XFree86 is one of the many implementations. True, XFree86 is reaching it's end of life. BUT that's not because it's a complete mess, because it's not, but it's due to the recent licence change. Indeed, the building process should be moved to a more GNU-like method (autoconf, automake, etc).

Indeed some of the new extensions do break a few things, and that's because translucent windows was inconceivable (is my spelling correct?) 10 years ago. There are solutions for all of them, some are hacks indeed because having an app written for X 15 years ago and still running perfectly is a reason of pride not of shame. Some apps just don't need modification.

It's great that X works even on 2 colour displays. It means that it will run also on a monocrome LCD. Cheap and efficient.

5) Nothing is easy. If it does so many things, in so many ways, on so many systems, with security and other features X has, it's great. Of course greatness comes at the cost of complexity.

The part with xine, well, that problem comes from the window-managers. The thing with the screen-saver was NOT an ugly hack, it just was the easiest way around it. YOU CAN CONFIGURE a screen-saver very easily, I'm not sure on how that is done from a program, but I can investigate.

Why do people doubt something that worked and works even on my 386 and on my dual xeon?

Y! is a cute thing. It tries to implement some things which are cute. But it gives up many of the things that make me respect X11. If it wouldn't then it would be just another X11 like software. Another thing: X has almost 20 years of programing behind it, they cannot beat that in short-term. They want to do what X and GTK (or QT or something else) do. Those are huge monsters. They are the basis of GUI, and have evolved incredibly. No matter how good they are, and how many they are, it's a gigantic task.