Supporting Xgl

[Taku]

Hi all,

I've always found very strange that Xgl server project was so little encouraged, whereas it's clearly the base of the future of Linux Desktop.

Ubuntu is now a great community, famous and respected, and its users ( us ) are so numerous that they ( we ) can now have a great influence in the open-source community ( an influence that came up with the Breezy release ),
so why not to participate, encourage or support the Xgl project.
Indeed this one need money, peaple, and, overall love ( one of the two main dev has left recently because he felt that nobody was interested in his work, especially because Exa, whereas Exa was intented to be a temporary solution for 2D graphic acceleration ).

So please, all, considert this message : it is time for ubuntu to make a strong and very significant act for the community, and supporting Xgl should be in my opinion the right one.

All together, let us make the significant step closer to the most technically advanced OS that linux must become.

[Benjamin_Lebsanft]

isn't Xgl the project which only has one active coder ? :rolleyes:

[JGJones]

I thought Xgl was replaced with the Xegl project?

One Google Job later...

Xegl:
http://www.freedesktop.org/wiki/Xegl

Xgl:
http://www.freedesktop.org/wiki/Software_2fXgl

[Taku]

isn't Xgl the project which only has one active coder ? :rolleyes:

That's exactly why we should support this project.

Only one active coder for the project which could change all the Linux desktop ...
We're wasting a change to put linux upper.

[macleod199]

I thought Xgl was replaced with the Xegl project?

One Google Job later...

Xegl:
http://www.freedesktop.org/wiki/Xegl

Xgl:
http://www.freedesktop.org/wiki/Software_2fXgl

I'm pretty sure Xgl is a larger piece that requires Xegl.

[poofyhairguy]

Hi all,

I've always found very strange that Xgl server project was so little encouraged, whereas it's clearly the base of the future of Linux Desktop.

Ubuntu is now a great community, famous and respected, and its users ( us ) are so numerous that they ( we ) can now have a great influence in the open-source community ( an influence that came up with the Breezy release ),
so why not to participate, encourage or support the Xgl project.
Indeed this one need money, peaple, and, overall love ( one of the two main dev has left recently because he felt that nobody was interested in his work, especially because Exa, whereas Exa was intented to be a temporary solution for 2D graphic acceleration ).

So please, all, considert this message : it is time for ubuntu to make a strong and very significant act for the community, and supporting Xgl should be in my opinion the right one.

All together, let us make the significant step closer to the most technically advanced OS that linux must become.

I would like this as well, but it seems that Exa is the direction many things are going. The problem is that Xgl will take a LONG time to mature compared to EXA and all the big Linux Desktop people want the "bandaid" out there quicker so that we at least have SOMETHING when Vista hits. Even with sponsership from Ubuntu, EXA will still win on the time scale so it should be what Ubuntu uses if Mark really wants to make a stand against Vista when its released he will put his eggs in the EXA basket. And neither solution solve the problem that besides the closed source Nvidia driver and a few Open Source ATI and Intel drivers, video card drivers for Linux suck. That is a big problem that needs a lot of band aid to patch until companies (aka ATI) support Linux better.

Plus, the person to hire for this would be Jon Smirl (they person who recenlty quit the project), and he has openly clashed with some important Ubuntu developers. So....well....at least SUSE has hired the only person working on it.

[Lovechild]

Actually Novell hired Dave Reveman not SuSE, he works with the Novell labs on Xgl for the next version of NLD.

And if we are serious about accelerated graphics on Linux, there should be wider support for the development of the Open Graphics card - it's the only card that will offer specs that one can really develop stuff like xgl using, relying on closed drivers to do the work for us is a bad solution. And they have very little support for some reason, I'd like to see Ubuntu and other distributions work closer with these guys getting good support into X for their card. Maybe even recommending it with Ubuntu?

I was forced to buy an outdated r200 card to get decent X using open source drivers, this is the only card that currently runs xgl and xegl as far as I understand, because it's the fastest card we have specs for and thus can get a stable development platform on. Graphics on Linux is not going to get better without improving drivers first.

[poofyhairguy]

Actually Novell hired Dave Reveman not SuSE, he works with the Novell labs on Xgl for the next version of NLD.

Ooops. You are correct.


I was forced to buy an outdated r200 card to get decent X using open source drivers, this is the only card that currently runs xgl and xegl as far as I understand, because it's the fastest card we have specs for and thus can get a stable development platform on. Graphics on Linux is not going to get better without improving drivers first.

Yep, old ATI card by far have the best open driver support of any graphic card in Linux. Until more appear, the Linux desktop cannot advance forward (community leaders refuse- for good reason- to rely on the closed drivers like Nvidias). The problem is that right now the hardware companies have the most resources to throw at drivers. Till they play ball (open their drivers) the Linux desktop might lag behind.

Personally I went the other way. I reeently bought a 6600 GT just for its closed drivers just so I can run xcompmgr. I know is a "toy" that has not been developed since 2004, but it might be the only way to get Vista and OSX like effects in the Gnome Desktop for some time. It works fine for me, Breezy fights with it much less than Hoary. Some can be patient, some can live with closed drivers and crashes. I'm in the second category.

[| MM |]

http://www.gnome.org/~seth/blog/xshots

This stuff looks promising, not sure if all that wobbly window stuff is what we want but.

But the workplace switcher/transparencies are very cool!

I hope it will make its way to Ubuntu.

Giving windows momentum or something so you can slide/flick them from workspace to workspace would be cool, hehe.

[poofyhairguy]

http://www.gnome.org/~seth/blog/xshots

This stuff looks promising, not sure if all that wobbly window stuff is what we want but.

But the workplace switcher/transparencies are very cool!

I hope it will make its way to Ubuntu.


It will. But I must tell you after playing with it that it eats a LOT of CPU. A lot lot. Just to move Windows. So until a framework is in place and we have some better graphics drivers, Luminocity will stay a toy.

[Rotarychainsaw]

What would Xgl do though? Don't programs have to be written to take advantage of whatever Xgl does? I mean we have cairo and glitz (right?) in breezy and I haven't noticed THAT much a difference. I know the clearlooks theme is being rewritten as we speak to use cairo.

[andrego]

ATI is never going to build drivers for Xgl if nobody is using it, and nobody will be using it because it's not done. Hence, I agree with Taku - we should support Xgl.

However, EXA is also interesting (Especially for older hardware, or other architectures not supported by NVidia!).

Having both would mean:

EXA = compatible acceleration for pretty much all systems
Xgl = awesome 3D effects and acceleration for anyone with a supported 3D accelerator card and supported drivers.

So, supporting EXA in the short-term probably isn't a bad idea at all. It will mean almost everyone having an accelerated desktop. Unfortunately it means we wait longer for some of the cooler features, but we're going to be waiting for a while anyway...the drivers for nVidia cards aren't even 100% compatible yet.

[Taku]

ATI is never going to build drivers for Xgl if nobody is using it
[...]
So, supporting EXA in the short-term probably isn't a bad idea at all.


I totally agree, Exa is on the way, we'll have, certainly before the end of the year the possibility to play with it, because the release of Xorg7.0, so in my opinion there is no special reason to "support" Exa, I mean that Exa is important, sure, but already a part of an *active* project.
It is assumed that supporting Xgl won't give us exceptionnal perspectives until a couple of years maybe, but it would be a very interesting long term *investment*.

Graphics on Linux is not going to get better without improving drivers first.

You're totally right, I think and hope that in very few time ( maybe already now ? ), ubuntu and its community will be strong and numerous enought to be able to get in touch with great companies like NVidia and ATI.
Dealing with them would be a big challenge, but in my opinion, Ubuntu is growing so fast that NVidia and ATI could find the way of the reason in some time ...
At least I'm dreaming of this ... Ubuntu shipped with ATI and Nvidia drivers out of the box.

( In a more realistic way, why not to support more the teams working on open-source drivers ... ? Ubuntu weight now enought to do this. )

Let's make Ubuntu better than all users could dream it.

[MadMan2k]

right, exa is the short term solution and xegl is the long term one.
Luminocity is the research project from Metacity and will be backported some time.

[apoclypse]

It surprises me that with so many cutting edge ideas, coming from the linux community, that X is still even around. Its old, very old, and has only seen anything significant change in the last few years. Xegl/gl I think would have been the real change. But like I'm finding out now where I work, it seems that change is not always welcome. The linux developement can be a stodgy one regardless of how much cutting edge ideas are out there, these usually stay just that, ideas with a proof of concept and no more. Exa is not the way in my opinion Xgl/egl is, it is the real solution to the problem, not only in terms of driver support, but in terms of being ahead the competition. Yes, Xegl will take time but it's time worth spending to ensure that we don't end up scrambling to play catch up like we are now (which is really what exa is all about) with the other desktops. This is whats happening now and if the developers were smart they try to avoid the same fate in the future.

Anyone who can help should help, it is sorely needed. If enough developers show interest, namely the ubuntu developers, and this community I'm pretty sure that the others will come around (they will have to, to save face).

[poofyhairguy]

You're totally right, I think and hope that in very few time ( maybe already now ? ), ubuntu and its community will be strong and numerous enought to be able to get in touch with great companies like NVidia and ATI.
Dealing with them would be a big challenge, but in my opinion, Ubuntu is growing so fast that NVidia and ATI could find the way of the reason in some time ...
At least I'm dreaming of this ... Ubuntu shipped with ATI and Nvidia drivers out of the box.

That is a dream. You need some backround.

I have an Nvidia card and it has great closed driver (I think). Do you know why? Its not because of me or my buying habbits or your either. The only reason Nvidia and to some extent ATI create Linux drivers AT ALL is because many studios are starting to use Linux and things like Blender to render 3D animation. They make the drivers to sell the high end cards. In fact, this explains why Nvidia gives better support- they were always better at OpenGL than ATI so because Linux and its high end software uses Opengl to do the work Nvidia started out with an advantage in the market.

I can almost promise you that if the high end card market did not exist or if these studios were not switching to Linux for rendering support from both companies would be next to nothing. Both really don't care about us- the desktop Linux user. ATI is improving their drivers to help sell high end cards, not to allow Linux users to play Doom 3 faster.

We have no control over them. None. The situation as it is puts them in complete control. If Ubuntu went to them and asked for something based on Ubuntu's desktop user base, both companies would run the representitives out with laughs. They really don't care about the Linux Desktop market that much.

So what we have is the best we can get for now. I'm greatful blender exists as it does because I know thats why I have good Nvidia drivers. I don't pretend. Until we get 10% or more of the desktop marketshare they have no reason to listen to our desktop demands. None. Thats the core of the problem.


( In a more realistic way, why not to support more the teams working on open-source drivers ... ? Ubuntu weight now enought to do this. )


Ubuntu is much smaller than you think. About 15 or so full time developers work on Ubuntu- I hear Nvidia has just that many or more in charge of their *nix drivers. Making drivers is a hard business expecially when the companies won't help. Progress is being made, but Ubuntu has bigger things to worry about.


When you start to try to figure out why the Linux Desktop is what it is it can be depressing. Its slow moving, and it will pretty much always lag behind the other major desktops. It makes sense when you figure out that now Linux is a server OS so none of the really big players (IBM, Google, etc.) that are involved with Linux have no reason to help develop the Linux desktop. They don't even really see a market for it besides hobbiests for some time- thats why there isn't a Desktop IBM OS based on Linux, and why Novell and Red Hat push Linux as a server more than a desktop. Almost no one believes in a Linux desktop at the top of the corporate world, and thats where the real money is. The 10 million Mark gave to Ubuntu is chicken scratch compared to what IBM invests in the server side of Linux. A fraction. And thats what we have to deal with.

There is a reason that the Xegl developer- Jon Smirl- quit: his project was not getting nearly enough of the support that he expected it to get. Why? Because none of the big players believe with all their hearts in the Linux desktop. Not yet at least. Ubuntu is not enough to change this tide- the organization uses all of its resources just getting a stable Sid released every six months....

[Taku]

You're right poofyhairguy ... this ins't realistic for now to hope that nvidia or ATI will tell us "ok guys, let's rock ... ", but what I said was that we must think more about long term developement, and encouraging people in developing basic and fundamental parts of linux is something that we should't discuss about.
It is just basic.

Linux community is, in my opinion, very often not looking far enought ...

Both Linux and Ubuntu community are growing fast, so when we'll be able to take advantage of this growth, and by the way to require better and maybe free drivers to ATI and Nvidia, what will happen ?

We'll be saying "Yes ! great day .... so maybe we should work again on a Xgl server now ...".

We have to make an effort to see farther, because Linux and Ubuntu are things in which we believe.

[GazaM]

poofyhairguy, that was a very pessimistic view and I have to seriously disagree... A lot of key players in Linux including Red Hat and Suse DO believe in desktop linux, the fact is that being realistic it will take time, but to say it will never be done because no companies care about desktop linux is misinformed... you being a respected member of these forums should know that the linux community is a development force to be reckoned with. It is the reason linux is where it is today, when one developer needs to let go of a project there are more to step in and continue on. Windows has bugs galore and the only reason people settle is because they have no choice, but linux gives us choice and human nature makes us take that choice for granted and notice more problems when compared to 'other' OS's. In fact OsX probably has the best desktop out of all, but I very rarely use Apple computers and so would not have enough experience to comment on that.

Another thing... Blender is NOT the single reason we have ATI and nVidia linux drivers today, I use Blender a lot and think it's great, but it is the multi-million dollar programs like Maya and XSI which hollywood studios use that make companies want to have drivers... it was definitely a contributing factor to their interest but studios are not the only thing that made ati and nVidia step up.... the linux desktop market is growing greater by the day, and with those linux supporting companies like id and the Unreal team linux is being recognised as a rising OS, and as market share increases game companies will follow suit, this is inevitable. Ati deserve special mention for recently having a surge in driver development for linux, with each driver bringing better features, performance and compatibility... there has also recently been a fix for ati cards going into the windows drivers which brings ATI's OpenGL performance way up, making the latest card rise above nVidias 7800GTX, and this should be implemented into the linux drivers, albeit in a while :P

Back to the point, the future of desktop linux is looking VERY promising, so dont put a downer on it, realise that we're not alone and as the consumer starts getting more aware of alternatives linux marketshare and therefore hardware and software support will rise :D

[poofyhairguy]

poofyhairguy, that was a very pessimistic view and I have to seriously disagree... A lot of key players in Linux including Red Hat and Suse DO believe in desktop linux, the fact is that being realistic it will take time, but to say it will never be done because no companies care about desktop linux is misinformed...

I won't say it will never be done. I do think that it will lag behind the other two established desktops. If I said it was impossible for something to ever be done, then I would be proved a fool. I just think its not worth holding your breath for.


you being a respected member of these forums should know that the linux community is a development force to be reckoned with. It is the reason linux is where it is today, when one developer needs to let go of a project there are more to step in and continue on.

One thing I understand is that investments focused on development of Linux's desktop side is a fraction of the investments focused on development of the server side. That changes over time, but its a waiting game at this point.


Another thing... Blender is NOT the single reason we have ATI and nVidia linux drivers today, I use Blender a lot and think it's great, but it is the multi-million dollar programs like Maya and XSI which hollywood studios use that make companies want to have drivers... it was definitely a contributing factor to their interest but studios are not the only thing that made ati and nVidia step up....

I guess we both don't really know. I was trying to use blender as an example- and everything I have seen shows me that Nvidia and ATI invest what they do in Linux because of the high end market. But we can agree to disagree on that one.


the linux desktop market is growing greater by the day, and with those linux supporting companies like id and the Unreal team linux is being recognised as a rising OS, and as market share increases game companies will follow suit, this is inevitable. Ati deserve special mention for recently having a surge in driver development for linux, with each driver bringing better features, performance and compatibility... there has also recently been a fix for ati cards going into the windows drivers which brings ATI's OpenGL performance way up, making the latest card rise above nVidias 7800GTX, and this should be implemented into the linux drivers, albeit in a while :P

I guess they got tired of the firegls not competing in the market.


Back to the point, the future of desktop linux is looking VERY promising, so dont put a downer on it, realise that we're not alone and as the consumer starts getting more aware of alternatives linux marketshare and therefore hardware and software support will rise :D

If I am pessimistic its because I do not beleive that Linux can keep up with OSX and eventually Vista when it comes to giving us a modern desktop. One day it will arrive and it will be nice, but the last thing I want anyone to do is expect the revolution to come soon. If Linux has a stable desktop as advanced as OSXs is now in late 2007 I will be happy. Things are moving faster with Xorg, but faster just means that things are moving unlike before.

http://dri.freedesktop.org/~jonsmirl/graphics.html


Conclusion

My experience with the failure of Xegl has taught me that building a graphics subsystem is a large and complicated job, far too large for one or two people to tackle. As a whole, the X.org community barely has enough resources to build a single server. Splitting these resources over many paths only results in piles of half finished projects. I know developers prefer working on whatever interests them, but given the resources available to X.org, this approach will not yield a new server or even a fully-competitive desktop based on the old server in the near term.


I don't 100% agree with Mr. Smirl, but I will say that his article shows what ALL that needs to happen- and he assumes the whole time that the driver issues don't exist. You add those to his analysis and it would make anyone be skeptical.

I can deal with it. I can chill with my xcompgr till the Linux desktop is ready. Many old school nerds hate to say that eye candy makes a difference, but to most users the small things are what matters most. Xorg will get there one day, and the drivers are getting better on the OSS side.

[Taku]

I disagree with your "general advice" poofyhairguy, but not really in the content of what you say.

More precisly, I think that wait the succes to invest more to create an opportunity of a bigger ( and really GREAT ) succes is a valid excuse only for those who don't really believe in our community.

I hope I did express myself correctly and that I was understandable, but I just wanted to say that we mustn't wait to succeed in our "Desktop contest" to begin to develop a basic and very big piece of the Desktop.

[poofyhairguy]

I disagree with your "general advice" poofyhairguy, but not really in the content of what you say.

More precisly, I think that wait the succes to invest more to create an opportunity of a bigger ( and really GREAT ) succes is a valid excuse only for those who don't really believe in our community.

Low blow.

Ok you are right. I don't believe in the community. Thats why I wrote the most complete composite manager guide on the internet:

http://www.ubuntuforums.org/showthread.php?t=75527

And thats why I begged the developers to not compile Breezy's Metacity with its compositor so that xcompmgr would be stable enough for use. Thats why I have helped many people here and on other forums improve their linux desktop to be the best it can be. Thats why I lead the tweaker team.

In fact, if I was going to make an insulting comment like you just did, I would say that those that makes demands of others to spend their money on the developers needed to finish certian projects are the ones who don't "believe" in the desktop Linux community. I would say that those people would need to do more themselves and quit asking things of others. But I won't make that claim, because I know the truth.


I hope I did express myself correctly and that I was understandable, but I just wanted to say that we mustn't wait to succeed in our "Desktop contest" to begin to develop a basic and very big piece of the Desktop.

Ok. I'm going to put this to rest. If Ubuntu did get another developer to work on Xgl , the only person qualified to do the work is Jon Smirl who worked on the Xegl. Mr. Smirl in the past has openly conflicted with some of Ubuntu's lead developers. So if Mark hired Mr. Smirl, he would be inviting problems into his team's "chemistry." No good boss would do that. Xgl has a future, and maybe Ubuntu will work on it more one day, but for now its all about EXAs.

[Taku]

Please poofyhairguy, don't take this for you, really, it was the message which I quickly wrote, not to make you a *dark heretic* but just to encourage people to think farther, to keep in mind that linux has got great potentialities and that we have to think "long term" ...

I know you're one of the most active user on this forum, and one of the most devoted member of this community, sincerely. So please don't take this message for you as an insult.

[poofyhairguy]

Please poofyhairguy, don't take this for you, really, it was the message which I quickly wrote, not to make you a *dark heretic* but just to encourage people to think farther, to keep in mind that linux has got great potentialities and that we have to think "long term" ...

I know you're one of the most active user on this forum, and one of the most devoted member of this community, sincerely. So please don't take this message for you as an insult.


Deal. No harm done.

[ltmon]

Hi Poofyhairguy,

I think that you *were* right about the original market for nVidia *nix drivers -- high-end animation studios.

There is evidence that both ATI and nVidia are caring more about the desktop market though. See this thread in the KDE panel-devel list, which describes ATi (and later nVidia) contacting the developers to be included in beta releases so that they can be sure their hardware is supported.

http://mail.kde.org/pipermail/panel-devel/2005-August/000435.html

Cheers,

L.

[aariel]

I suppose I'm a little late for this conversation, but I just finished compiling what they have of xgl and I'm wondering if any of you have done the same and can offer some insight regarding its inner workings =] sorry for the off-topic, but this seems like it's in the interest of "Supporting Xgl"

[Anthem]

I'm pretty curious to see what happens with the new open graphics card. I'll probably shell out for one just to support the project, but that looks like it has the potential to be pretty cool.

[poofyhairguy]

Hi Poofyhairguy,

I think that you *were* right about the original market for nVidia *nix drivers -- high-end animation studios.

There is evidence that both ATI and nVidia are caring more about the desktop market though. See this thread in the KDE panel-devel list, which describes ATi (and later nVidia) contacting the developers to be included in beta releases so that they can be sure their hardware is supported.

http://mail.kde.org/pipermail/panel-devel/2005-August/000435.html

Cheers,

L.

Thats good news.

[wbeck85]

Does anyone know about the r300 project over at sourceforge (http://r300.sourceforge.net/)? Has anyone dared to try the drivers on their Radeon card? Could it be an ATI fix to Xgl? or anything related?

[drizek]

poofy is right. the drivers are not meant for desktop cards. they are designed for the workstation ones.

think about it, fglrx. atis workstation cards are called FireGL.

[MadMan2k]

poofy is right. the drivers are not meant for desktop cards. they are designed for the workstation ones.

think about it, fglrx. atis workstation cards are called FireGL.
thats where the driver came from - but since the early days the driver development crew changed, the driver appeared on Ati's Homepage and is now listed for every Ati Card >= Radeon.

[pardasaniman]

It's not just animation studios.. Alot of imaging and hardcore processing is being done on gigantic non-windows clusters these days, often one needs alot of GPU horsepower to visualize the data.

[wbeck85]

I'm pretty curious to see what happens with the new open graphics card. I'll probably shell out for one just to support the project, but that looks like it has the potential to be pretty cool.


What do you mean by "new open graphics card?" who is making it? and you mean that all the drivers are open source? any links?

[GeneralZod]

What do you mean by "new open graphics card?" who is making it? and you mean that all the drivers are open source? any links?

http://lists.duskglow.com/mailman/listinfo/open-graphics
http://kerneltrap.org/node/5743

I'd also be happy to buy one just as a show of support, although whether the first few iterations would be worth actually using remains to be seen.

[poofyhairguy]

thats where the driver came from - but since the early days the driver development crew changed, the driver appeared on Ati's Homepage and is now listed for every Ati Card >= Radeon.

But I would argue that in the weekly meeting, the reason the little boss can convince the big boss at ATI to keep spending money on the drivers is for the high end market.

It just so happens the high end market is just super powered versions as the low end market, so its not that hard to make the drivers work for the weaker cards.

[poofyhairguy]

I'd also be happy to buy one just as a show of support, although whether the first few iterations would be worth actually using remains to be seen.

Problem is the card has to be better than a 9200. Who knows if they can do that!

[MadMan2k]

But I would argue that in the weekly meeting, the reason the little boss can convince the big boss at ATI to keep spending money on the drivers is for the high end market.
I would reply that the bigger issue with Linux support is not the programming itself but the customer care, if you want to do it right.
And Ati offers customer care for everyone with a readon+, which surley are in majority not the rendering studios.

[merlyn]

Hi Poofyhairguy,

I think that you *were* right about the original market for nVidia *nix drivers -- high-end animation studios.

There is evidence that both ATI and nVidia are caring more about the desktop market though. See this thread in the KDE panel-devel list, which describes ATi (and later nVidia) contacting the developers to be included in beta releases so that they can be sure their hardware is supported.

http://mail.kde.org/pipermail/panel-devel/2005-August/000435.html

Cheers,

L.

Perhaps Nvidia & ATI, would care more about the desktop market if more games were available for Linux.

Apart from the big rendering studios, gaming on PC would have to be an important market.

Am I barking up the wrong tree here?

[23meg]

Finally, the day has come when people started figuring out that we need open source hardware in addition to open source software.

[Norm 2782]

Finally, the day has come when people started figuring out that we need open source hardware in addition to open source software.

I really don't think we need that that much... Don't get me wrong, it would be great. But if Windows can get the max performance out of a closed-source driver, Linux should be able to do that too. Of course IF there are good performing and well documented drivers. And that's the problem unfortunately. Ubuntu is paving the way for a great Linux desktop (and Windows replacement) for all (nerds and noobs alike) and when that time is there, nVidia and ATI WILL support Linux as much as Windows. But we have a long way to go before we get there.

[GeneralZod]

Problem is the card has to be better than a 9200. Who knows if they can do that!

Indeed so. I'll probably just buy one of the initial ones as a way of providing funding (the guys in charge don't want to take donations for various reasons) and use it is as a paperweight :)

[darius779]

Does anyone know about the r300 project over at sourceforge (http://r300.sourceforge.net/)? Has anyone dared to try the drivers on their Radeon card? Could it be an ATI fix to Xgl? or anything related?

I dont think the r300 project at sourceforge is still active. The r300 driver has been integrated into the dri project (dri.freedesktop.org) and ships (will ship?) with X.org. It has nothing to do with ATI, and nothing specifically to do with xgl as IIRC xgl will require new drivers all around (which is one of the near term disadvantages of xgl)

The r300 driver works right now, though its performance isn't really wonderful, and as far as I know, it doesnt yet have EXA support. Still, it does work and performance will only get better so Im excited to try it out in xorg 7.0 and get rid of fglrx.

[wbeck85]

Oh. Ok, thanks for the info

[kamstrup]

Guys; the succes of the Linux desktop is not solely dependant on top notch graphics drivers...

Right now I'm hacking on the deskbar project (http://raphael.slinckx.net/deskbar/) which will hopefully be accepted into Gnome for 2.14. I can tell you that when we hit 1.0, Steve Jobs of Apple fame will wan't our heads on a plate.

Deskbar is similar to Spotlight (http://www.apple.com/macosx/features/spotlight/) in many ways, except:
- It is faster
- It's more flexible

You guys have no idea how fast beagle is! Best is ridiculously slow - Beagle is a monster. With the bleeding edge deskbar (not in cvs yet) I get all my search results in < 0.5s (on a a 1400MHz Duron!) Oh, and it's querying on the fly ofcourse...

There I go fanboying my own app :) ... But, man is it going to rock.

[bonega]

Kamstrup: not to sound negative, but gnome will not include any mono apps in the base packages.
Ubuntu might include it(really don't know their take on mono).

[tim1]

Kamstrup: not to sound negative, but gnome will not include any mono apps in the base packages.
Ubuntu might include it(really don't know their take on mono).

You obviously didn't really read the website. There are many more data sources than just beagle, but if you have beagle installed, it's an additional plus. And it says also om the website that the integration is done via libbeagle, which is a c lib, no mono in there.

greets, tim

[poofyhairguy]

Guys; the succes of the Linux desktop is not solely dependant on top notch graphics drivers...

Right now I'm hacking on the deskbar project (http://raphael.slinckx.net/deskbar/) which will hopefully be accepted into Gnome for 2.14. I can tell you that when we hit 1.0, Steve Jobs of Apple fame will wan't our heads on a plate.

Deskbar is similar to Spotlight (http://www.apple.com/macosx/features/spotlight/) in many ways, except:
- It is faster
- It's more flexible

You guys have no idea how fast beagle is! Best is ridiculously slow - Beagle is a monster. With the bleeding edge deskbar (not in cvs yet) I get all my search results in < 0.5s (on a a 1400MHz Duron!) Oh, and it's querying on the fly ofcourse...

There I go fanboying my own app :) ... But, man is it going to rock.


That looks awesome. I agree, eye candy is not everything.

[Lovechild]

something about deskbar rubs me the wrong way, I watched the flash demo a while back and.. it just seems wrong but it's hard to put my finger on it.

I think it's to much functionality folded into a commandline like application, it doesn't seem gnomish in it's basic idea - I can type to start an app, to search google, to search beagle - I mean it sounds useful but.. something about it just doesn't seem right.

[quickgun]

Getting OT here, but to me, it's essentially a persistent run command dialog on steroids.

It could possibly serve as a great unifying usability app for the gnome environment, but we'll have to wait and see.

BTW:
Does it/will it/shouldn't it use d-bus?

[ember]

Mhmm ... I basically have the same feeling ... so much functionality conveys an impression of 'either that will never be stable or it will only work with xyz' to me ...
It is a cool idea nevertheless.

[super]

ya the deskbar thing looks cool but i agree with this comment by lovechild: I think it's to much functionality folded into a commandline like application,

this (to me anyhow) seems like it should be more than a gnomepanel app almost like it would be underused if it was just in the panel. i picture this being more like a permanent fixture on the desktop. kinda like a sidebar gdesklet but better.

[Lorkki]

ya the deskbar thing looks cool but i agree with this comment by lovechild:

this (to me anyhow) seems like it should be more than a gnomepanel app almost like it would be underused if it was just in the panel. i picture this being more like a permanent fixture on the desktop. kinda like a sidebar gdesklet but better.

At the very least the panel is always visible, even if you like working with apps in maximised mode like I do. To access some desktop panel/applet I'd have to unmaximise the windows, always retain one empty virtual desktop, or make said extra panel/applet hog up extraneous screen space. I would at least very much like to retain a normal panel applet as an option in such a case.

[bonega]

You obviously didn't really read the website. There are many more data sources than just beagle, but if you have beagle installed, it's an additional plus. And it says also om the website that the integration is done via libbeagle, which is a c lib, no mono in there.

greets, tim

I see - thanks for correcting me.
Seems like I have to check it out

[kamstrup]

Just to clear things out; deskbar is written in Python. It uses 100% user controllable pluggins, so you are not forced to search for everything... You can scale it down to mini-commander functionality if you wan't.

Lovechild: The flash demo is outdated... To the point of being completely misleading.

Well, sorry to hijack the thread :-)

[kamstrup]

On-topic-ish

Graphics is not (and has never been) one of the strong sides of the Linux desktop. But then ask the question, what is it good at?

Linux is widely regarded as a server OS. Why not use that to buff the desktop then... In some way this is exactly what beagle does. Linux/GNU/X/Gnome is stable... this is also a plus on the desktop. Continue this list your selves...

We don't have accelerated graphics on the desktop, but we still have more eye-candy than windows (XP)...

[poofyhairguy]

We don't have accelerated graphics on the desktop

We don't have STABLE accelerated graphics. I use xcompmgr everyday.

[daniels]

Xgl will be in universe. Maybe Luminocity too.

[kamstrup]

Wow! Cool. I was going to say that you guys rock Daniel, but I guess you hear that often - so I wont :)

[Taku]

What a great news !!

Really Daniel ? not a joke ?

I'm so excited ... and I just can't hide it ... tututu ...

Really ... what a great news ... please keep us informed about this.

[daniels]

http://lists.ubuntu.com/archives/dapper-changes/2005-October/000250.html

[MadMan2k]

anyone tried this yet? will it replace sxer-xorg as default, or will I have to start it manually?
Does it even make sense to try it out, i.e. will I get some hardware accellaration with my graphics card?

ok, did you know Google?
$ ./Xglx :1 -ac -screen 640x480

but how do I let apps connect to this Server?

[Norm 2782]

anyone tried this yet? will it replace sxer-xorg as default, or will I have to start it manually?
Does it even make sense to try it out, i.e. will I get some hardware accellaration with my graphics card?

ok, did you know Google?
$ ./Xglx :1 -ac -screen 640x480

but how do I let apps connect to this Server?

Best thing I managed to achive is :
xterm -display :1

Should put a xterm in your Xgl server. From there you can launch other programs. Gonna check if I can get a window manager running tonight.

[LichiMan]

Will Xgl be the server by default in Dapper?

[GeneralZod]

Will Xgl be the server by default in Dapper?

Very unlikely, alas - not only is XGL itself not ready for primetime (nor likely to be within the next year), but the graphics-card driver situation under Linux will mean that XGL is usable only by a minority.

[GazaM]

... but the graphics-card driver situation under Linux will mean that XGL is usable only by a minority.

This whole thing about graphics drivers reminds me a lot about the multimedia codecs issue... people seem to forget that Windows doesnt include proprietry 3d drivers out of the box either, the drivers must be installed seperately, EXACTLY like it is with linux, in fact linux is easier as they are in the repos...

The best solution IMO would be to have Xgl there, but have it disabled by default... that way people can enable it after they have set up their drivers etc. Or have the system auto-detect whether you have hardware acceleration or not and enable it based on that.

[daniels]

Xglx will absolutely not be the default; it's too much of a pain.

[yfkar]

This whole thing about graphics drivers reminds me a lot about the multimedia codecs issue... people seem to forget that Windows doesnt include proprietry 3d drivers out of the box either, the drivers must be installed seperately, EXACTLY like it is with linux, in fact linux is easier as they are in the repos...
I was going to say that the difference is that we don't have (for example) very stable 3d drivers for newer ATI cards but my Ubuntu hardlocked while I was writing. ;)

[MadMan2k]

Best thing I managed to achive is :
xterm -display :1

Should put a xterm in your Xgl server. From there you can launch other programs. Gonna check if I can get a window manager running tonight.
ok, you can start all the gtk apps with the following:

<app-name> --display=:1 &

if you start nautilus, metacity and the gnome-settings-daemon, you will get an nice nested server - but wheres the exciting stuff?
There must be a way to try xcompmgr and such...

[Anthem]

This whole thing about graphics drivers reminds me a lot about the multimedia codecs issue... people seem to forget that Windows doesnt include proprietry 3d drivers out of the box either, the drivers must be installed seperately, EXACTLY like it is with linux, in fact linux is easier as they are in the repos...

That's not true. Sorry.

Windows includes proprietary ATI and nVidia drivers on the windows CD.

[GazaM]

That's not true. Sorry.

Windows includes proprietary ATI and nVidia drivers on the windows CD.

Anthem... maybe my Windows cd is faulty then, because I didn't get any proper proprietry ATI or nVidia drivers with it (and my cd is 100% original, NOT warez). There are drivers however in Windows Update but not on the install disc for me. The first thing people do after they install Windows is insert the cd's which came with the hardware to install drivers, including graphics card drivers.

Perhaps it depends on how recent the card is (mine is a PCI-Express X700Pro), maybe the drivers are just very old, because I did not get any with it.

[MadMan2k]

Perhaps it depends on how recent the card is (mine is a PCI-Express X700Pro), maybe the drivers are just very old, because I did not get any with it.
thats it - my R9800 is being recognized by WinXP sp2 and some old ati driver is being used.

[poofyhairguy]

The best solution IMO would be to have Xgl there, but have it disabled by default... that way people can enable it after they have set up their drivers etc. Or have the system auto-detect whether you have hardware acceleration or not and enable it based on that.

As of right now, I think the only cards that accerate Xgl is the ATI 9200. Not many people have them.

[Anthem]

Yeah, it has to do with how old the cards are. My GeForce4 was "recent" when my WinXP cd was mastered, and it's got proprietary drivers on there. Same with the ATI integrated chipset back when I ran Windows on this laptop.

If the CD included drivers for hardware more recent than the CD itself, then Microsoft's obviously been funnelling R&D money to more projects than just software... :D

[super]

This whole thing about graphics drivers reminds me a lot about the multimedia codecs issue... people seem to forget that Windows doesnt include proprietry 3d drivers out of the box either, the drivers must be installed seperately, EXACTLY like it is with linux, in fact linux is easier as they are in the repos...

mmm, not really.
with windows you can get hardware acceleration with almost every graphics card available (even if you have to download it yourself). in linux you can only get it if you have an nvidia or an ati card.

and the number of cards that support hardware acceleration under the xgl server is even smaller.

unfortunately...

[daniels]

in linux you can only get it if you have an nvidia or an ati card.
Or SiS, or Intel i8xx/i9xx (which, by the way, ship by far the most chips out of *any* vendor), or Matrox, or S3/Via, or Trident, or 3dfx. Or Glint, or SunFFB, or ...

[Anthem]

Or SiS, or Intel i8xx/i9xx (which, by the way, ship by far the most chips out of *any* vendor), or Matrox, or S3/Via, or Trident, or 3dfx. Or Glint, or SunFFB, or ...

Really? I've never seen SiS give proper 3D rendering.

[kamstrup]

Or SiS, or Intel i8xx/i9xx (which, by the way, ship by far the most chips out of *any* vendor), or Matrox, or S3/Via, or Trident, or 3dfx. Or Glint, or SunFFB, or ...

In theory yes, but just because X supports it doesn't mean that the kernel/rest of stack does... Example (https://bugzilla.ubuntu.com/show_bug.cgi?id=15954).

Cheers

[clash]

I think that this should be given priority for the next version of Ubuntu.

[leech]

Anthem... maybe my Windows cd is faulty then, because I didn't get any proper proprietry ATI or nVidia drivers with it (and my cd is 100% original, NOT warez). There are drivers however in Windows Update but not on the install disc for me. The first thing people do after they install Windows is insert the cd's which came with the hardware to install drivers, including graphics card drivers.

Perhaps it depends on how recent the card is (mine is a PCI-Express X700Pro), maybe the drivers are just very old, because I did not get any with it.

No, you're right, unless the card is extremely old (like the Matrox G400, which DID have Direct3D drivers out of the box for WinXP, but it did NOT have OpenGL...) The 3D drivers that were included with WinXP were very basic all around. They don't include even all the functions that the cards of the day could handle.

Leech

[macleod199]

No, you're right, unless the card is extremely old (like the Matrox G400, which DID have Direct3D drivers out of the box for WinXP, but it did NOT have OpenGL...) The 3D drivers that were included with WinXP were very basic all around. They don't include even all the functions that the cards of the day could handle.

Leech

Yeah, this burns me every time I re-install Windows 2000. Especially as somehow they don't support my Network Card (with the ever standard on Linux 8139 chipset). Makes things difficult. :)

[nocturn]

Poofy is right, the state of Linux graphic drivers is sad.

I have a machine with a SiS M760 chipset which has a 3D engine, but no drivers.
So the framerate is stuck at 300fps... nothing I can do about it :-(

[poofyhairguy]

Poofy is right, the state of Linux graphic drivers is sad.

I have a machine with a SiS M760 chipset which has a 3D engine, but no drivers.
So the framerate is stuck at 300fps... nothing I can do about it :-(

My mom has that chipset in her computer and I can't even do 3ddesktop with it.

No way it could handle Xgl or EXA.

[kamstrup]

Yeah, this burns me every time I re-install Windows 2000. Especially as somehow they don't support my Network Card (with the ever standard on Linux 8139 chipset). Makes things difficult. :)

Hehe yeah. Reminds me of when I installed XP for my mom in law (she uses Ubuntu now ofcourse...). The network card just refused to work. I called the shop where we bought the computer... "Well. You have to download drivers from our homepage...".

Sweet irony. Sorry OT; I know. Just had to share ;-P

[nocturn]

My mom has that chipset in her computer and I can't even do 3ddesktop with it.

No way it could handle Xgl or EXA.

Sucks, doesn't it... I hope they get it reverse engineered or SiS eventually releases specs...

[seguso]

Hi all,
Indeed this one need money,

If someone sets up a "group action" on Fundable (http://www.fundable.org), I will gladly put 25 or 30 euros on it. ;)

[zAo]

Hehe yeah. Reminds me of when I installed XP for my mom in law (she uses Ubuntu now ofcourse...). The network card just refused to work. I called the shop where we bought the computer... "Well. You have to download drivers from our homepage...".

Sweet irony. Sorry OT; I know. Just had to share ;-P

Like the old BIOS-'bug': remove your keyboard and boot up, your BIOS will say something like:
"Keyboard error, press F1 to continue" :cool: