Any news on Plymouth?

I see that there's a Gobby document for Plymouth, but it's empty! Does anyone know if there's news from UDS on Plymouth in Ubuntu?

Well i hope we at least get this since we're not getting Login Experience!

Does noone know anything about this? :confused:

I guess it would mean a great deal of work to get Plymouth working 100%, so perhaps since noone has heard anything it's not going to happen for 9.04. I'll still keep my hopes high though ;-)

It's not gonna happen for 9.04 but we may get a PPAs for people that are interested in trying it out and have the supported hardware. Agreement is to have it by default for 9.10 when all the drivers are ready and all problems are solved. Here is the video of the session where you can see the decision process in more detail: http://videos.ubuntu.com/uds/jaunty/Desktop/2008-12-11/morning-post-break/00001.ogv

Thanks :-) Exactly the kind of information I was looking for.

It might still be available for those of us who wish to test it out eventually, hopefully. I am kind of hoping to have something quite representable such as facebrowser or plymouth in. Those two, new default theme, quick boot, and we're all set ;) well, of course, a few other things... video related issues... audio related issues... etc... hmm... -__-; Dri2, Ext4, and all the other goodies are obviously welcome :)

Jay_Bee said: It's not gonna happen for 9.04 but we may get a PPAs for people that are interested in trying it out and have the supported hardware.

Agreement is to have it by default for 9.10 when all the drivers are ready and all problems are solved. Here is the video of the session where you can see the decision process in more detail: http://videos.ubuntu.com/uds/jaunty/Desktop/2008-12-11/morning-post-break/00001.ogv

I tried looking at the video but as don't know any of the players (speakers and contributors) its hard to know who said what. Without some reference its meaningless :(

It's not really necessary that you know who is who, you can get that from what they are talking about. I don't know them either, I recognized Mark S. sitting in white shirt, asking few questions, he asks about including it in Jaunty and gets the response that it's not recommended. Then he talks about the possibility of a PPA for 9.04. The man sitting next to him is obviously someone from the Foundations Team. Also, the man with deep voice talking at the beginning is Ken Wimer, head of the Artwork Team... Hope that helps...

Any news on the PPAs for this? Hopefully we'll get PPAs for Login Experience too.

Jay_Bee said: It's not gonna happen for 9.04 but we may get a PPAs for people that are interested in trying it out and have the supported hardware.

Agreement is to have it by default for 9.10 when all the drivers are ready and all problems are solved. Here is the video of the session where you can see the decision process in more detail: http://videos.ubuntu.com/uds/jaunty/Desktop/2008-12-11/morning-post-break/00001.ogv

ah you have got to love Ubuntu's commitment to free driver work, even extending it to testing. I love how that is not left up to other people such as the Fedora community. oh wait...

the fedora community should stop bitching. oh wait... it's just you.

Agreed, You always come here and brag about fedora and how Ubuntu does nothing.....Hmmm, I think Opensuse had the most commits this year on KDE4 AND STILL SUPPORTED KDE3, Which Fedora just openly dropped. Fedora is great for testing out whats up and coming in the next few months, and thats what fedora and redhat advertise it as. Meanwhile other distros like to be fairly stable or at least try to stay on the development level with everyone else. Case and point is opensuse and KDE4 network manager, all done by novell and opensuse.

quickshade said: Agreed, You always come here and brag about fedora and how Ubuntu does nothing.....Hmmm, I think Opensuse had the most commits this year on KDE4 AND STILL SUPPORTED KDE3, Which Fedora just openly dropped. Fedora is great for testing out whats up and coming in the next few months, and thats what fedora and redhat advertise it as. Meanwhile other distros like to be fairly stable or at least try to stay on the development level with everyone else. Case and point is opensuse and KDE4 network manager, all done by novell and opensuse.

I think you both misunderstood, this is being kept out because it's not considered stable enough (and to be honest it probably isn't, lots of work to be done). The problem naturally becomes who does the work, if everybody adopts a wait and let everyone else do the development and testing attitude we will all slow to a crawl and never again will stability happen.

Look at Intrepid, it ships with OpenOffice 2.x precisely because of this relutance to import beta/release candidate work for testing early in the development cycle. The final 3.0 certainly didn't become a least bit more stable by not being exposed to users. In the end this type of decision making ends up not serving anyone.

It worries me that we have to rely on other people to do our testing and development. Most of all what worries me is that Ubuntu might get a little to comfortable at it's current pole position. I've been around for a while, Ubuntu is certainly not the first media darling Linux. The only ones with real staying power have been ones who stepped up to the plate and helped improve the platform.

Yes I am involved with Fedora but I also have Ubuntu running (Intrepid currently but as I am rather underwhelmed by it I plan to jump on Jaunty soon). I have to support Ubuntu users, that also means I take an interest in how Ubuntu is developed. Fedora does a lot of great things, it doesn't mean they should be alone in doing it. If we want influence on where Linux goes, the way forward is participation. If we keep not adopting early and helping out, they get to decide the future. Is that in Ubuntu' best interest?

It is also about building expertise, at some point Canonical has to make money. Currently it does not, to be able to fulfill service contracts one needs to be able to fix bugs. Getting aggressive with regards to the platform is the way to build that expertise. Paying consulting fees to get bugs fixed sounds unappealing as an approach to turning red numbers black, and red numbers have a tendency to stop the adventure at some point.

Even if one was to consider Canonical "the competition", I want to see them prosper. By them having success, everyone benefits. They might invest a large part of their future profits in the open and free Linux platform we all use and love. Regardless of my personal choice of distro, that is a good thing for me. Money made by Canonical is more likely to be money taken from Apple and Microsoft not to mention money from growth in before undiscovered IT areas - like netbooks and MID, which a year ago barely existed - than money taken from existing Linux businesses, some will be but I wager not a lot.

Buttomline: It's about the future. Do we want influence and an assured way to build a profitable business, then we need to get our fingers dirty and make sure progress happens. I hope we want that, it's good fun on top of everything.

I think part of the lack of adventurousness is that ubuntu is commited to having a (reasonably) stable and useable release every 6 months rather than a release when ready model,this leads to a more conservative approach since they have to "guess" 6 months in advance what can be done in that time frame . While those of us who are testing junkies might like a little more daring, a rush to the bleeding edge would not fit Ubuntus "model". Should all distros ( atleast the "major" ones) contribute to developement ? Yes but all of us holding hands and marching off a cliff would not serve Linux/opensource well.Note I am not saying Fedora is "marching off a cliff" that is just "poetic liscense".

Well i have to agree with Gnomeuser and Ronacc here. Ubuntu focuses on stability more than bleeding edge but we have not seen that stability in the last two releases or much bleeding edge. I also agree that we need to test more Beta and RC stuff. The 2.6.28 Kernel is hardened for Ext4 and we have been able to use the 2.6.28 Kernel for a couple of weeks now but Ext4 is not available in the installer. If Grub2 is made default which was discussed at UDS, they should have released 2 versions of Jaunty Alpha 2, one with Grub2 and one with Legacy. If they decided to stay with Legacy for Jaunty Final then the people who prefered or want to stick with Grub2 can do so for testing purposes.

Login Experience should be put in the PPAs ASAP so that even though it is not gonna be in Jaunty, those of us that want to test can begin doing so now and start filing bugs and pave the way for it to be used in 9.10.

As soon as we get KMS the same should happen with Plymouth. I do agree that Plymouth is a little young for Jaunty though specially since it needs Kernel 2.6.29 which wont happen for a while.

OOo-3 should have been in Jaunty since Alpha 1.???

Packagekit isnt default yet.(If it is going to be in Jaunty which it looks like)

FF-3.1 should already be default. Its very stable and works well. Jaunty is gonna release with it anyway.

Just because Ubuntu decides not to go with a feature for an upcoming release doesnt mean it shouldnt be available to those who would like to test it and maybe have been looking forward to it. This increases the testing time by a large amount too and ensures that whats being tested now will be rock solid for the next LTS.

Jaunty will be fine with or without <insert feature other than theme overhaul>.

the problem is of course that to many things need to be done to get this to work, kernel modsetting, which is unlikely to happen as of right now, Xorg changes which will be in the next release which at this point is unknown when that will be. Then you have the video drivers that need to be released, After all that you still have the problem of mixing all these into one and seeing what works and what doesn't. I agree that we need to test these or they will never be stable, but at the same time with no real release schedules for some of these how can you include it in a distro (and build around it) only to find out that 3-4 of the other things you need won't be done by feature freeze. Not really time well spent. This was the case with opensuse 10.3 and KDE4, they wanted to include 11.0 in 10.3, but didn't think that the release schedules would allow it, and postponed it. Turned out to be great because 4.0 was a "development" version and needed a lot of work (see 4.1) to make it generally usable. Now we are at a point where they are going for a 9 month release so that they can really help development on 4.3.

All in all what needs to happen is a test version of Ubuntu that will never be converted to stable that includes all these changes, With only the installer being stable. 1/4 of the people that post here say they always run the unstable versions so why not have one that is always just a test version for up and coming features in Ubuntu.

what you mean is like the unstable repository that Debian has. There was some talk on doing that couple of releases ago but then the idea was dropped. Not enough contributors and testers perhaps.

Starks said: Jaunty will be fine with or without <insert feature other than theme overhaul>.

Why other than theme overhaul?

Slug71 said: Why other than theme overhaul?

It is fairly likely that he does not understand that to do a complete overhaul of the look and feel of the platform, we need these underlying features to be in place.

Take the new notifications e.g., that beautiful transparency depends on composition. To get that you need 3D accelerated graphics. to get in any kind of supportable way you need DRI2, KMS and Gallium3D. This foundation then allows proper accelerated drivers which we can support.

Till that work is in place one would have to shortcut to getting that support or have a fallback that does not depend onf compositing. Both are undesirable from a support and QA stance. The latter solution includes the problem that it likely will not look or perform as nice as the composited solution.

People think overhauling the theme is such a simple job, just take a mockup and turn it into reality. There is a lot of work that needs to go into that process.

gnomeuser said:

Take the new notifications e.g., that beautiful transparency depends on composition. To get that you need 3D accelerated graphics. to get in any kind of supportable way you need DRI2, KMS and Gallium3D. This foundation then allows proper accelerated drivers which we can support.

Well... maybe its Omega 10 you are talking about...:confused:

ftp://ftp.infradead.org/pub/spins/README.Omega-10-Release-Notes

Release a heavily broken distribution... insane IMHO but maybe possible for Fedora :confused: Just double moral in cubic !

ShirishAg75 said: what you mean is like the unstable repository that Debian has. There was some talk on doing that couple of releases ago but then the idea was dropped. Not enough contributors and testers perhaps.

Yep, but keep it semi stable. Like how opensuse offers the lastest snapshots (about once a week) of KDE unstable. Great for testing before beta releases and RC releases and a good way to judge feature sets. Ubuntu could easily do the same thing with the whole OS being updated with new builds maybe every 2 weeks and then let testers have at it, once the features mature enough they are put into a pool that will be dumped into the next release.

Example: Plymouth and kernel 2.6.29 and the lastest version of Xorg and all released on this test repo. Bugs are reported to the developers of these programs (not ubuntu developers themselves, unless it's a packaging error) Once they become stable enough for a wide release (or beta testing) they are moved over into the next version of ubuntu (so lets say it takes 2 months to stabilize a set of features, it's moved into the fall release of ubuntu, to allow more time for testing.) If some gets put in here (maybe a new version of networkmanager) and within a month it's fairly stable, someone can suggest it get moved up to the next release and not pushed back to the fall release as it seems fairly stable.

This addresses 2 issues we have right now, it makes testing all these apps easier for users (just one repo and even live cds) and developers don't have to push as hard to get people to test new features, It also makes it easier to determine if something can make a current release (hey we tested amarok 2.1 and it's not really that buggy and think it can make the spring release, or if it is just not ready (plymouth.)

quickshade said: All in all what needs to happen is a test version of Ubuntu that will never be converted to stable that includes all these changes, With only the installer being stable. 1/4 of the people that post here say they always run the unstable versions so why not have one that is always just a test version for up and coming features in Ubuntu.

quickshade said: Yep, but keep it semi stable. Like how opensuse offers the lastest snapshots (about once a week) of KDE unstable. Great for testing before beta releases and RC releases and a good way to judge feature sets. Ubuntu could easily do the same thing with the whole OS being updated with new builds maybe every 2 weeks and then let testers have at it, once the features mature enough they are put into a pool that will be dumped into the next release.

Example: Plymouth and kernel 2.6.29 and the lastest version of Xorg and all released on this test repo. Bugs are reported to the developers of these programs (not ubuntu developers themselves, unless it's a packaging error) Once they become stable enough for a wide release (or beta testing) they are moved over into the next version of ubuntu (so lets say it takes 2 months to stabilize a set of features, it's moved into the fall release of ubuntu, to allow more time for testing.) If some gets put in here (maybe a new version of networkmanager) and within a month it's fairly stable, someone can suggest it get moved up to the next release and not pushed back to the fall release as it seems fairly stable.

This addresses 2 issues we have right now, it makes testing all these apps easier for users (just one repo and even live cds) and developers don't have to push as hard to get people to test new features, It also makes it easier to determine if something can make a current release (hey we tested amarok 2.1 and it's not really that buggy and think it can make the spring release, or if it is just not ready (plymouth.)

Exactly what i was trying to say here.

Slug71 said: If Grub2 is made default which was discussed at UDS, they should have released 2 versions of Jaunty Alpha 2, one with Grub2 and one with Legacy.

If they decided to stay with Legacy for Jaunty Final then the people who prefered or want to stick with Grub2 can do so for testing purposes.

It would be the best way to marry Stability and Bleeding Edge.

Also bringing in new features around Alpha 3/4 is too late IMO. If they have been tested and are good or a release is stable like OOo-3 or Ext4 then they should be in by Alpha 1/2. The more time goes on the less testing time there is.

Also like Plun has said a few times, maybe its time we stop waiting on Debian all the time.

In order to do this we need to talk to the developers about this, and this doesn't have to be "ubuntu only" You could easily have other distros jump in.The whole point of this would be to have a whole bunch of people testing unstable features and reporting bugs on them, and if you start getting 3-4 distros doing the same thing bug reports are going to be duplicated and the testing area to widespread. One distro that developers can requests their programs be dropped into and once they are fairly stable they are deemed "suitable for main distro testing" and are dropped into the different pools of distro test versions. One thing this would really help out is the whole release schedule issue, as you don't have to beg developers from Ubuntu or fedora to pick up your program for their next version, just submit it to this distro, get it approved and then it gets dropped in and everyone test it from there. Much better than some distros passing and some accepting it and then finding out it's not stable enough and dropping it or having another distro picking it up later because it is stable enough.

quickshade said: In order to do this we need to talk to the developers about this, and this doesn't have to be "ubuntu only" You could easily have other distros jump in.The whole point of this would be to have a whole bunch of people testing unstable features and reporting bugs on them, and if you start getting 3-4 distros doing the same thing bug reports are going to be duplicated and the testing area to widespread. One distro that developers can requests their programs be dropped into and once they are fairly stable they are deemed "suitable for main distro testing" and are dropped into the different pools of distro test versions. One thing this would really help out is the whole release schedule issue, as you don't have to beg developers from Ubuntu or fedora to pick up your program for their next version, just submit it to this distro, get it approved and then it gets dropped in and everyone test it from there. Much better than some distros passing and some accepting it and then finding out it's not stable enough and dropping it or having another distro picking it up later because it is stable enough.

Yep, This will also greatly help improve integration of new versions of KDE, Gnome, Xfce and OpenOffice which maybe contain big new features and is used across the board and is also where Packagekit will be important since it is also likely to be used by many distros soon. The same may even be said with Plymouth once KMS is ready.

Maybe time for an Ubuntu Brainstorm? Though that didnt go to well with Login Experience or Plymouth.

quickshade, this is exactly what mark shuttleworth was talking about sometime back . Google up "Mark Shuttleworth syncronicity" some links for you.

http://www.markshuttleworth.com/archives/150 https://wiki.ubuntu.com/MarkShuttleworth

If I remember correctly he wanted to sync everyones development schedules to release in the fall and spring, while that might be a good idea, this is opensource and people shouldn't be bound to a release window just because the distros "force" them to. My idea is more along the lines of "heres a distro you can ask to submit you project to and get it tested, once it's stable enough we will release it to each distros testing version. So in reality it would be like one giant alpha version distro with new software coming in and out often. Allowing better testing with the people who always want to run unstable, but not from svn or something. I'd be willing to help out with this, but we will need to get some developers from a couple distros on board to do this.

quickshade said: If I remember correctly he wanted to sync everyones development schedules to release in the fall and spring, while that might be a good idea, this is opensource and people shouldn't be bound to a release window just because the distros "force" them to. My idea is more along the lines of "heres a distro you can ask to submit you project to and get it tested, once it's stable enough we will release it to each distros testing version. So in reality it would be like one giant alpha version distro with new software coming in and out often. Allowing better testing with the people who always want to run unstable, but not from svn or something. I'd be willing to help out with this, but we will need to get some developers from a couple distros on board to do this.

I really like where this is going.

Gnomeuser would be the person on here to talk too about getting Fedora in on this.

Slug71 said: I really like where this is going.

Gnomeuser would be the person on here to talk too about getting Fedora in on this.

Not really, you would have to approach the Fedora Board to get the schedule changed. That being said Fedora and Ubuntu are aligned pretty nicely, Ubuntu is generally out a couple of weeks before Fedora.

The problem Shuttleworth is wanting to attack is that he doesn't have the manpower to pull of a proper enterprise ready LTS. It is thus in his best interest to get Novell and Red Hat to pull much of the weight. He envisions this happening by aligning all the major enterprise vendors on the same schedule. By doing this and agreeing to use the same basic kernel, toolchain etc. everyone would have a reduced support burden.

Now why would Red Hat and Novell do this, they already carry the majority of the burden doing backports and testing for their enterprise releases. They have their own tools and workflows built up around their existing solutions. They would gain very little short term and be giving up a whole lot of freedom to schedule (not to mention it would rapidly cut into their profits).

Don't get me wrong long term more cooperation is definitely a good thing, but I would love to see a proposal like this come with an implementation in Free Software of the testing framework Mark envisions. Some kind of prototype, call it a gesture of good will if want, something to show that Canonical are serious and want to commit to this. Giving us a forum for even doing would be wonderful. This will not happen ovenight, and if it is ever to happen we need someone to sit down and build the tools in an open fashion. I think Canonical could do that quite well. Right now though, it's seeming to everyone who can help make this happen like Shuttleworth using his massive media presence to pressure everyone to do his bidding without him investing a dime. Everyone needs to feel like equal partners or it will never get off the ground, I believe his vision is good long term but his approach to making it happen is fatally miscalculated.

I imagine that there might be a vanue to start a kind of OpenEnterprise.org companion to FreeDesktop.org, to create specs and tools that suit the enterprise. Tools we can collaborate around. I think that would be the way to unify us all a bit more around our common goals. Figure out what problems we need to solve and sitting down, solving them together as partners.

It being Christmas time and all I'm going to be gone for the next 2-3 days, As I said though if we can get some developers to help us set up a way to build this and advertise it as a test distro that would be a good way to start. I think it's one of those ideas that really needs people to see how great of an idea it is and how helpful it will be before they will devote more time to it. So if 2-3 Ubuntu developers can sit down and help us get a plan laid out and then we can present it to all the developers, they might be more willing to commit, maybe even produce a prototype.

gnomeuser said: Not really, you would have to approach the Fedora Board to get the schedule changed. That being said Fedora and Ubuntu are aligned pretty nicely, Ubuntu is generally out a couple of weeks before Fedora.

The problem Shuttleworth is wanting to attack is that he doesn't have the manpower to pull of a proper enterprise ready LTS. It is thus in his best interest to get Novell and Red Hat to pull much of the weight. He envisions this happening by aligning all the major enterprise vendors on the same schedule. By doing this and agreeing to use the same basic kernel, toolchain etc. everyone would have a reduced support burden.

Now why would Red Hat and Novell do this, they already carry the majority of the burden doing backports and testing for their enterprise releases. They have their own tools and workflows built up around their existing solutions. They would gain very little short term and be giving up a whole lot of freedom to schedule (not to mention it would rapidly cut into their profits).

Don't get me wrong long term more cooperation is definitely a good thing, but I would love to see a proposal like this come with an implementation in Free Software of the testing framework Mark envisions. Some kind of prototype, call it a gesture of good will if want, something to show that Canonical are serious and want to commit to this. Giving us a forum for even doing would be wonderful. This will not happen ovenight, and if it is ever to happen we need someone to sit down and build the tools in an open fashion. I think Canonical could do that quite well. Right now though, it's seeming to everyone who can help make this happen like Shuttleworth using his massive media presence to pressure everyone to do his bidding without him investing a dime. Everyone needs to feel like equal partners or it will never get off the ground, I believe his vision is good long term but his approach to making it happen is fatally miscalculated.

I imagine that there might be a vanue to start a kind of OpenEnterprise.org companion to FreeDesktop.org, to create specs and tools that suit the enterprise. Tools we can collaborate around. I think that would be the way to unify us all a bit more around our common goals. Figure out what problems we need to solve and sitting down, solving them together as partners.

I didnt mean to speak to you about aligning the releases but more about a bleeding edge alternate ISO which pretty much stays a testing platform by which stuff get tested and once good to go it gets moved into the main release cycle. If you get what im saying. Read the last few post between mine and Quickshades and you might get a better idea. I think if Say Fedora, Ubuntu, Mandriva and Opensuse did something like this, stuff like the new(RC/Beta/Alpha) versions of Gnome, KDE, OO, Packagekit, Plymouth, FF, Ext4, Grub2 etc, All stuff which gets used across the board will get much better testing and over a longer timeframe. For example we could now be testing Login Experience on this test release but its only coming to Ubuntu in 9.10. Grub2, Packagekit and FF 3.1 could also be default in this version. If we had something like this since the dev cycle of Intrepid we could have been toying with OO-3 already but now theres all these build problems and we're already approaching Alpha3 of Jaunty and we still dont have it. May as well skip it now and go for 3.0.1 which would already be in this test dev ISO had we had one.

The synchd release cycles would be nice to though but i do understand the complications that it could bring. Good idea to start with a Forum though like you mentioned.

quickshade said: It being Christmas time and all I'm going to be gone for the next 2-3 days, As I said though if we can get some developers to help us set up a way to build this and advertise it as a test distro that would be a good way to start. I think it's one of those ideas that really needs people to see how great of an idea it is and how helpful it will be before they will devote more time to it. So if 2-3 Ubuntu developers can sit down and help us get a plan laid out and then we can present it to all the developers, they might be more willing to commit, maybe even produce a prototype.

Yep, ill Brainstorm it for now.

[[IMG]http://brainstorm.ubuntu.com/idea/16814/image/1/[/IMG]](http://brainstorm.ubuntu.com/idea/16814/)

Slug71 said: Yep, ill Brainstorm it for now.

Great way to start, I think I'll bring this up over at the opensuse camp, see what they say. [[IMG]http://brainstorm.ubuntu.com/idea/16814/image/1/[/IMG]](http://brainstorm.ubuntu.com/idea/16814/)