[00:59] <ScottL> ailo_at, are you still working on the theme as well?
[01:23] <ScottL> ailo_at, would you mind looking at bug #720476
[01:24] <ScottL> and marking it as "confirmed" if you are experiencing the same problem
[01:24] <ailo_at> ScottL: Was it hard to file the bug? I confirmed it.
[01:25] <ailo_at> I mean, I marked that it affects me as well..
[01:26] <ailo_at> Ok, I confirmed it.
[01:36] <ScottL> ailo_at, filing it?  no, wasn't hard at all actually :)   i used the command that persia suggest "ubuntu-bug debian-installer"
[01:36] <ScottL> it absolutely rocked
[01:37] <ailo_at> ScottL: I didn't look through the attachments, but did you see the cause for it at all?
[01:37] <ScottL> it started collecting information, opened a new tab in firefox, started the bug report, and attached the "relevant" information
[01:38] <ScottL> ailo_at, i didn't look through it really, there were many attachments and i would be looking for a needle in a haystack because of my ignorance
[01:38] <ScottL> but i'm hoping to poke someone tomorrow at work in #ubuntu-installer as persia suggested
[01:39] <ailo_at> I suppose someone will find out, if the info is missing from the bug report.
[01:39] <ScottL> i say "relevant" because it may not contain any information from the failed install at all!  especially after backing up and not selecting the video-meta and completing the install :/
[01:39] <ScottL> but there's a bug report now and i'll poke around in #ubuntu-installer and maybe the can suggest a better way to find out
[01:41] <ailo_at> ScottL: I've been working on the -controls the past two days. Still learning the basics. I copy what I can from the original controls
[01:41] <ailo_at> But using the gui I made.
[01:42] <ScottL> it looks like you and paultag are working together, no?
[01:42] <ScottL> which is an awesome way to learn...he told me he was part of the beginner team that helps people learn
[01:43] <paultag> yeah, I've been really sick, I'm just on now
[01:43] <paultag> ScottL: ailo_at rocks, he's got a ton done
[01:43] <ailo_at> I had a problem yesterday with hidden tabs (symbols that you don't see from a text editor), that comes from copy and pasting from the web. :P
[01:44] <ailo_at> Spent a couple of hours scratching my head
[01:44] <paultag> ailo_at: heheh
[01:47] <ailo_at> paultag: Just pushed an update on the Glade file. The code is still just a barebone. The only thing that is partly worked on is the memlock control, and the help file opens the ubuntustudio docu wiki page.
[01:47] <ScottL> paultag, heh, i was hella sick yesterday and just getting over it today, think it was a slight bit of food poisoning :/
[01:48] <paultag> ailo_at: awesome
[01:48] <paultag> ScottL: oh no :(
[01:48] <ScottL> oh, ailo_at and paultag: not that we should stress on this now, but someone had a really good suggestion :
[01:48] <ScottL> have a graphical way to "upgrade" to ubuntu studio from ubuntu desktop
[01:49] <ScottL> that might fit into -controls as some point, just a single click "upgrade" button to do it all
[01:49] <ScottL> just thinking out loud, maybe this isn't the place
[01:49] <ailo_at> ScottL: I have a button called "Set UbuntuStudio Default Settings" that is supposed to do that.
[01:49] <ScottL> and certainly not for this release but i'll add it to the "release planning"...
[01:50] <ScottL> oRLy?
[01:50] <ScottL> well, that unequivocally rocks ailo_at :)
[01:51] <ailo_at> ScottL: Allthough, it doesn't install audio packages at all, but I suppose we could expand that a little. Add metapackage selection to the controls.
[01:51] <ScottL> now that i have an install of ubuntu studio on my second computer that seems stable i can start testing kernels as well
[01:51] <ScottL> ailo_at, don't stress on it yet, that can be a feature to be added later for sure
[01:51] <paultag> :)
[01:51] <ailo_at> Adding meta packages should be easy enough
[01:58] <ailo_at> Another thing I've thought about is having an indicator app for the controls. The app we are working on now could be a system preferences part of an indicator menu.
[02:05] <ScottL> i've been making good use of my latest Linux Format magazine today while incapacitated and have been reading jono's effusing on unity and i realized...
[02:05] <ScottL> when i installed studio today it went straight to gnome instead of unity!
[02:05] <ScottL> so the patch i did actually worked!  yay :)
[02:05] <ScottL> thanks TheMuso for uploading it :)
[02:05] <ScottL> and i'll have to send didirock a big thank you tomorrow as well for patiently working with me to wrangle up the patch
[02:06] <ailo_at> I have to start installing from network. I'm running out of DVD's :P
[02:06] <ScottL> ailo_at, i've been using rewritable dvd's, they're good for 20 or so rewrites before they start acting up
[02:07] <ailo_at> I usually install from usb on the computers that support it. Would be nice to get rid of the CD DVD format all together.
[02:07] <ScottL> i wonder however, etiquette for bugs, can i mark the gnome-classic bug as "fixed" now, or should i wait until the official natty release?
[02:08] <ScottL> and i need to report this bug fixed on the team report also :)
[02:28] <ScottL> holstein has been pushing for a graphical installer, which i believe will then allow you to create usb images to install
[02:28] <ScottL> maybe in a non-lts version we can implement the graphical installer and just install ALL the packages for that release
[02:29] <ScottL> and work on a way to select (if we feel we still need to do select packages*) packages for the next release
[02:29] <ScottL> this would give us a year to work on implementing package selection in a graphical installation
[02:30] <ScottL> * i suppose it is possible to decide not to select packages if we have a streamlined package selection anyway
[02:30] <ScottL> i noticed the 'audio production' menu fits on the screen currently in natty, which is a good thing in my opinion :)
[02:32] <ailo_at> Not for me, since I have a 1024x768 resolution.
[03:04] <ailo_at> ScottL: Just thinking about the graphical installer. Isn't it better to have a meta package for that in Software Center?
[03:05] <ailo_at> The controls can still have package selection, at least meta-package selection.
[03:05] <ailo_at> And the way you would install from the -controls would be: choose packages, hit apply.
[03:18] <ailo_at> If we made the controls into a indicator menu app, we could split up the functions. We could have a dedicated Multimedia package management app and a preferences app.
[03:18] <ailo_at> As menu items in the controls
[03:20] <ailo_at> falktx idea of a control app seemed very all in one. If the controls could dock other software into it, like jack, it could be similar to what falktx is working on.
[03:21] <ailo_at> I mean qjackctl, not jack itself
[03:21] <ailo_at> Could be nice to have a sound check function too.
[08:12] <abogani> ailo_at: Could you rerun tests disabling AUTOGROUP (that is do as root "echo 0 > /proc/sys/kernel/sched_autogroup_enabled"), please?
[08:17] <abogani> ailo_at: For find culprit you could use perf (with its function_graph tracer).
[12:40] <ailo_at> abogani: So, I installed linux-tools for 2.6.38-3 for now. Running perf gave this error "perf_2.6.38-3: error while loading shared libraries: libbfd-2.21-system.so: cannot open shared object file: No such file or directory". I have /usr/lib/libbfd-2.21.0-system.20110216.so, created today, so I made a symbolic link at /usr/lib/libbfd-2.21-system.co. It seems to work. You think this is ok to do?
[13:01] <abogani> ailo_at: It could be safe.
[13:01] <abogani> should*
[13:06] <abogani> ailo_at: Perhaps sched_switch is better trace...
[13:06] <abogani> But I'm really know...
[13:06] <abogani> *don't*
[13:06] <ailo_at> abogani: So much to look through. Could you give me an example command?
[13:07] <abogani> I know how find latencies culprit on realtime kernel but not on "normal" kernel :-(
[13:14] <abogani> As root
[13:14] <abogani> cd /sys/kernel/debug/tracing
[13:14] <abogani> echo 0 > tracing_enabled
[13:14] <abogani> cat availabe_tracers
[13:15] <abogani> echo tracer_choosen > current_tracer
[13:15] <abogani> echo 1 > tracing_enabled
[13:15] <abogani> do what do you want
[13:15] <abogani> echo 0 > tracing_enabled
[13:15] <abogani> cat trace
[13:16] <abogani> ailo_at: Do you have already made test with turning AUTOGROUP off?
[13:17] <ailo_at> I turned it off, but no change as far as I can see. No reboot needed, right?
[13:17] <abogani> No.
[13:17] <abogani> No reboot needed.
[13:19]  * abogani thinks that for do this works we should build a custom kernel...
[13:21] <ScottL> here is something a little surprising: http://ubuntu-tweak.com/app/ulatency/
[13:21] <ScottL> claims to lower latency by using cgroups
[13:21] <ScottL> abogani, i have done as you enumerated
[13:22] <ScottL> https://launchpad.net/~slavender/+archive/broken
[13:24] <abogani> ScottL: really good guy.
[13:24] <ScottL> ailo_at, just to make sure i understand your previous comments, are you suggesting that we create a singe "install ubuntu studio"meta package so that someone could upgrade by installing a single package via software center or synaptic or CLI ?
[13:24] <abogani> Studio still use alternate cd/dvd?
[13:24] <ScottL> abogani, yes
[13:24] <abogani> mhhhh
[13:25] <ScottL> ailo_at, i also like falktx idea of a single control app, i even suggested to him that he incorprate patchage or similar for connections in it as well
[13:25] <ScottL> ailo_at, this would really give a single point of usage to control the audio server
[13:25] <ScottL> abogani, but there are serious talks about moving to the graphical installer
[13:26] <ScottL> it's just a matter of understanding how and having the commitment from people to implement/maintain it
[13:26] <ScottL> and i suppose we need to address a few items as well, like if we continue to give people installation choice via tasksel and how to implement it
[13:26] <ailo_at> ScottL: I am for the single install package in Software Center. I'm not sure about having your own software doing the patchage and stuff, since there already are so many out there. I don't like putting everything in one program. Would be better to have it able to dock programs, like if they were plugins.
[13:27] <ailo_at> ScottL: An indicator menu like the one for sound and mail, maybe?
[13:28]  * abogani don't think that ulatencyd could help us. It seems a "replace" for AUTOGROUP (features added recently into the kernel also called "200 line magic patch" or similar)
[13:28] <abogani> thinks*
[13:29]  * ScottL is heading to work
[13:35] <abogani> Ok the next step for update lowlatency kernel is:
[13:35] <abogani> Repeat the same steps for ubuntu-natty-lowlatency (done for -meta).
[13:36] <abogani> *Before* launch debuild:
[13:36] <abogani> mkdir ~/a_temporary_directory_wherever_you_want -p
[13:36] <abogani> cp -a debian* ~/a_temporary_directory_wherever_you_want
[13:36] <abogani> cd ~/a_temporary_directory_wherever_you_want
[13:37] <abogani> and so debuild -S -sa 
[13:37] <abogani> and all the rest (like -meta)
[13:38] <abogani> ScottL: ^
[13:38] <abogani> It is simple, isn't it?
[13:38] <abogani> ;-)
[13:44] <persia> For those interested in graphical installers for the point of graphical installers, rather than for some other reason, plenty of work has been done at http://wiki.debian.org/DebianInstaller/GUI to provide a GUI for the current installer, which may be interesting, to avoid having to reengineer everything for a completely different installer.
[13:48] <abogani> We can't use the same Ubuntu installer?
[14:51] <persia> abogani, So, there are two different sorts of installer available for the architectures we support.  One is used on the live images (ubiquity), and the other is used on the alternate images (debian-installer).
[14:51] <persia> Most of the actual installation code is shared (it comes from other packages and the code is delivered within the context of the selected installer).
[14:51] <persia> Right now, Ubuntu Studio only has alternate images, so all the testing has been done with debian-installer.
[14:52] <persia> Adding a GUI to that is 90% testing: it might just work.  It might need a couple tweaks.
[14:52] <persia> Switching to ubiquity means changing the install experience, writing wrapper handlers for any installer components that aren't currently supported (e.g. for task handling), and then testing/tweaking.
[15:29]  * abogani admits that alternate install not suit well with Studio's audience ...
[15:31] <holstein> persia: i'll look at that link
[15:31] <holstein> i would just want to make sure that USB stick installs are simplified
[15:31] <holstein> which ever way we decide to go
[15:31] <holstein> *to try and cover as many issues as possible
[15:32] <holstein> i think adding the GUI to what we have would be cool
[15:32] <persia> I don't see any reason why USB installs shouldn't be completely independent of the technology used during the install.
[15:33] <holstein> well, currently there is an error
[15:33] <persia> But yeah, just making what we have use GTK rather than curses seems the line of least resistance to me (assuming the code works: I don't believe it's ever been tested in Ubuntu)
[15:33] <holstein> where the installer cant locate the media
[15:33] <holstein> thats using unetbootin
[15:34] <holstein> i havent tried the ubuntu USB stick maker in a while
[15:34] <holstein> but back at 10.04
[15:34] <holstein> it didnt do anything
[15:34] <holstein> also, having a LIVE installer would be nice
[15:34] <holstein> for testing purposes
[15:35] <holstein> so someone can fire up JACK and a synth on their hardware
[15:35] <holstein> for example
[15:35] <persia> I don't think we can get support for unetbootin at any point.
[15:36] <persia> I've heard there's work to port USB Creator to other platforms (Windows and OS X).  I'm unsure of current status.
[15:36] <holstein> ubiquity works with unet
[15:36] <holstein> and usb creator
[15:36] <holstein> thats why i was pushing for it
[15:36] <persia> I'm still opposed to LIVE, especially because if someone fires up JACK and a synth on their hardware in a live environment, they probably won't be able to play anything with anywhere close to acceptable latency.
[15:36] <holstein> seemed to take out the most birds with one stone ;)
[15:37] <holstein> yeah?
[15:37] <holstein> well, i was pushing for it for several reasons
[15:37] <holstein> it makes the installer look and feel like normal ubuntu
[15:37] <persia> Try it.  Boot a live Desktop CD, install qjackctl and a synth, and see what performance you get.
[15:37] <holstein> makes USB installs easier
[15:37] <persia> All prior testing found it completely unacceptable.
[15:38] <holstein> and
[15:38] <holstein> we are the only audio distro i can find that doesnt have a live installer
[15:38] <holstein> and my testing has been actually quite acceptable
[15:39] <holstein> the old dynebolic acutally gave a bit better performance live for me
[15:39] <persia> You've tried the test I described above?
[15:39] <holstein> that ubuntustudio 8.04
[15:39] <holstein> not a fair comparison
[15:39] <holstein> BUT
[15:39] <holstein> AVlinux live is quite similar to ubuntustudio installed
[15:39] <holstein> and KXstudio
[15:39] <holstein> not bad
[15:39] <holstein> do-able
[15:40] <holstein> persia: i'll try that test again in natty
[15:40] <holstein> just to make sure
[15:40] <persia> OK.  And no fair if you have really-new-extra-super-powered hardware :)
[15:41] <holstein> nope
[15:41] <persia> But if you think it's acceptable, maybe things have changed in the past few years.
[15:41] <holstein> dynebolic was running on a p3
[15:41] <holstein> with 256 ram
[15:41] <holstein> that was able to have jack stable at 20ms
[15:41] <holstein> persia: well, would i make an album on a live CD of ubuntustudio?
[15:41] <holstein> no
[15:42] <holstein> but, i think it would be handy
[15:42] <holstein> i think a lot of bad vibes occur for new users
[15:42] <holstein> for example
[15:42] <holstein> folks dont know that ubuntustudio = ubuntu
[15:42] <persia> I'm not opposed if it's acceptable.  I'm mostly opposed if it can give a poor impression of Ubuntu Studio.
[15:42] <holstein> so, they wipe their functional buntu install
[15:42] <holstein> OR dual boot
[15:42] <persia> We need to fix that.
[15:43] <holstein> persia: we should talk about this more
[15:43] <persia> We're Ubuntu Studio, but our difference to Ubuntu Desktop isn't enough to make us not Ubuntu.
[15:43] <holstein> because that is completely my motivation as well
[15:43] <holstein> all i can do is look at my experiences
[15:44] <holstein> and notice what i see in the support channel
[15:44] <holstein> for me, i installed 8.04
[15:44] <holstein> ubuntustudio
[15:44] <holstein> several times
[15:44] <holstein> didnt even know about the IRC back then
[15:44] <holstein> all i knew was my firepod didnt work
[15:45] <holstein> BUT, it worked with the 64studio live CD
[15:45] <holstein> and i didnt know why
[15:45] <holstein> it was simple permissions pretty much
[15:45] <holstein> BUT for me to test ubuntustudio
[15:45] <holstein> i had to install each time
[15:46] <holstein> and that at the time gave me a poor impression of ubuntustudio
[15:46] <persia> Aha.  Yes, that's a useful case for  a LiveCD.
[15:46] <persia> As long as stuff actually *works*.
[15:46] <holstein> and im arguing that this could be the case for others as well
[15:46] <holstein> however, i totally see your point of view as well
[15:46] <holstein> i feel like maybe we can come to some comfortable middle ground
[15:46] <persia> I'm mostly echoing the historical point of view.  In practice, I don't care that much.
[15:46] <holstein> and will address all of my concerns
[15:47] <holstein> and keep the performance in mind
[15:47] <persia> My main goal is to make sure we don't repeat mistakes because we've forgotten our history.
[15:47] <holstein> persia: i gotta split for a while
[15:47] <holstein> i'll catch up later though :)
[15:47] <persia> Sure :)
[16:02] <holstein> alright
[16:02] <holstein> couple more minutes
[16:02] <holstein> i sked cjwatson about the switch to ubiquity
[16:03] <holstein> hypothetically of course
[16:03] <holstein> i was told it would not be trival
[16:03] <holstein> BUT do-able
[16:03] <holstein> and the biggest loss for us is the metapackage selection
[16:03] <holstein> i would like to bring this up to though
[16:03] <holstein> i remember the first probably 6 or so times i saw it
[16:04] <holstein> i would hit the wrong key
[16:04] <holstein> and skip past it
[16:04] <holstein> accidentally
[16:04] <holstein> and not get any of them
[16:04] <holstein> and i should have RTFM'd probably
[16:04] <holstein> but i had not seen that kind of installer before
[16:05] <holstein> and didnt understand it fully
[16:05] <holstein> then, somewhere around 9.10
[16:05] <holstein> the metepackages wouldnt install
[16:05] <holstein> it took a while for me to figure out which one *not* to install
[16:05] <holstein> so that then install wouldnt fail
[16:06] <holstein> in this scenario, the end-user needed to figure out a work around for a bug
[16:07] <holstein> then, download a significant amount of data twice
[16:07] <holstein> just to finally see the ubuntustudio desktop
[16:07] <holstein> *and not know how to join a wifi network ;)
[16:08] <holstein> anyways... maybe this can be a time when we can really think about wht we want going forward
[16:08] <holstein> i mean, functionally, the alternate installer is fine
[16:08] <holstein> we all know that
[16:08] <holstein> but, i think we can all agree that something needs to be changed in the near-ish future
[16:09] <holstein> even if its just better USB stick installation documentation
[16:09] <holstein> scrap the live-cd idea
[16:09] <holstein> and a GUI installer kinda like the one we've got
[16:10] <holstein> i gotta run again, but
[16:10] <holstein> persia: are there any other options for installers that you are thinking about?
[16:10] <holstein> we got the current alternate
[16:10] <holstein> ubiquity
[16:10] <holstein> the debian GUI
[16:10] <persia> That's about it.
[16:11] <holstein> thats what i was thinking
[16:11] <persia> And the difference between the current one and the debian-GUI is presentation-only: no functional changes.
[16:11] <persia> Doing it in ubiquity is some functional work.
[16:11] <holstein> yeah, thats nice
[16:11] <holstein> just adding the GUI is easy
[16:11] <holstein> assuming the meta packages work*
[16:12] <holstein> well, we got options
[16:12] <holstein> maybe scott-work ScottL ailo_at abogani 
[16:12] <holstein> whoever
[16:12] <holstein> if you get a minute to read the scroll back
[16:12] <holstein> and comment or just think it over...
[16:13] <persia> Technically, that's task selection, not metapackage selection, although the names are the same, and the tasks all include the metapackages :)
[16:14] <holstein> good point
[16:14] <holstein> that helps me understand how one can be broken
[16:14] <holstein> and not the other
[16:15] <holstein> biab...
[16:41] <scott-work> sorry it took a bit, but i read the backscroll...exciting stuff :)
[16:41] <scott-work> work is kinda crazy today since i was out yesterday :/
[16:41] <scott-work> just to pull some other threads into this conversation...
[16:41] <scott-work> i was talking with ailo_at yesterday and i made myself question the task selection part
[16:42] <scott-work> if we have reduced the number of packages that are installed, perhaps we don't even need to worry about asking for user input during installation and just install all that we feel is necessary
[16:42] <scott-work> i imagine that users would find that a live cd /graphical installation is more important that being able to pick and choose between five task selections
[16:42]  * scott-work ducks and runs for cover now
[16:43] <persia> ScottL, With the GUI I referenced, we get task selection without extra work.  With ubiquity, we lose task selection, or we have to implement something complex (and that for natty+1, because feature freeze is RSN).
[16:44] <persia> Main thing to ask about task selection is if there are serious users who don't want the audio stuff.
[16:44] <persia> At one point there were some.  I don't know about now.
[17:05] <scott-work> persia: i think this might get to the core of *what* is ubuntu studio...both from what users expect and the reasons we make it
[17:05] <scott-work> i say that not to be flippant, i'm serious
[17:05] <persia> I don't take it flippantly, and I agree.
[17:05] <scott-work> i believe we don't do anything special for video or graphics that can't be done with synaptic, software center, or apt-get install
[17:05] <persia> But while I'm happy to discuss history and theory, I leave vision and decision to you.
[17:06] <persia> I believe we don't do anything special for audio that can't be done with a package manager.
[17:06]  * scott-work runs around screaming "We're doomed!  We're doomed!"
[17:06] <scott-work> kidding
[17:06] <persia> I further believe that if we *do* anything special, we're doing it wrong.
[17:06] <persia> The point of a flavour is to express an opinion about appropriate defaults and package selection.
[17:06] <scott-work> we adjust a few settings, like user in audio group and maybe including the -lowlatency kernel once it's in the repos
[17:07] <persia> Anyone can replicate the end-state with any package manager, *but* by having a flavour, people are saved the effort of doing so, and it Just Works.
[17:07] <scott-work> but this furthers my point: we would someone download a larger image, have to use a dvd, loose the live aspect of it, just to have gimp installed?
[17:08] <scott-work> when you can download a smaller, live cd, install it quicker with a fancy gui and less clicks and then add gimp?
[17:08] <scott-work> s/we/why
[17:10] <scott-work> i think our brand is audio and we tune the distro to it, we are even discussing a live dvd so that people can make sure their hardware works
[17:11] <ailo_at> I think puredyne has a nice combination. A live CD and a live DVD. You get both a slimmer version and a full version. DVD's can be tough to burn. USB will not work for everyone.
[17:11] <ailo_at> So, they cover all those problems
[17:12] <ailo_at> Main reason why I stopped installing US from DVD was I often got some corruption when burning, so it failed to install at some point
[17:12] <ailo_at> I'm checking out puredyne again to try performance.
[17:13] <ailo_at> It's 9.11, based on Karmic.
[17:14] <persia> If puredyne is based on Ubuntu, someone ought scavenge it for nifty patches :)
[17:14] <persia> ScottL, If the brand is audio, then yes, all the non-audio stuff doesn't matter at all.
[17:14] <persia> The interesting question is whether the brand is audio.
[17:17] <scott-work> persia: what was the demand previously for non-audio applications?
[17:17] <persia> There were a number of folk interested in video and video applications.
[17:18] <persia> And there was some work done to improve support for tablets, improve gimp/inkscape/etc. as tools, and otherwise ensure the quality of the experience for graphics production.
[17:18] <persia> I have no idea about user counts.
[17:19] <persia> Asking this a different way: who cares for firewire permissions, blender interfaces, tablet support, testing of inkscape/gimp/etc. if not Ubuntu Studio?
[17:20] <persia> I'm not saying we do this well today, but I wonder if it is determined to be out of scope if we can find someone else who will do this well.
[17:20] <scott-work> hmmm, good points
[17:21] <ailo_at> For myself I have very little knowledge about Video and Graphic applications, and if there are any problems setting up a system to work with those. 
[17:22] <ailo_at> Fonts, formats?
[17:22] <persia> So, setting up the system isn't especially hard.  Nothing special.
[17:22] <persia> There's really two areas that get interesting:
[17:22] <persia> 1) software: do the selected tools install/run smoothly.  Is someone watching their bugs, and making sure they get upstream, etc.
[17:23] <persia> 2) hardware: do we have working support for tablets, for jog:shuttle devices, for firewire cameras, etc.  Does RAW import and SANE do what we expect?
[17:24] <persia> firewire cameras were a dispute point for a long time.  With juju it should matter less, because it's now possible to have a solution, but I don't know if anyone is testing/tweaking this.
[17:25]  * persia has no firewire cameras, nor tablets
[17:35] <scott-work> i wonder how many of these improvements (like udev rules for firewire) have made their way into ubuntu desktop as well now, maybe by simply working with upstream
[17:35] <scott-work> ?
[17:39] <persia> Some.
[17:39] <ailo_at> Since a lot of multimedia/audio distros are based on Ubuntu and each one of them is tackling the problems with what apps to include (allthough they all are somewhat similar) I don't see why Ubuntu Studio must work for everyone. That is why I like the idea of a slim Live CD with only the essential stuff on it (some apps take very little space too).
[17:39] <persia> I know of several specific cases where the necessary stuff has just happened.
[17:40] <persia> But where I'm less sure is if there is anyone regularly testing: in the past we broke wacom support at least once during each release cycle.
[17:41] <persia> ailo_at, I don't think it's interesting to consider derivatives as real alternatives for users.  I instead consider them cases where we've failed to find a useful way to cooperate with someone.
[17:43] <persia> And I doubt I'm alone in wishing that if other folk are tackling the same question of which packages to include by default, they were doing so with us.  It's not easy to find the right answer, and users aren't well served by 10 different choices, each requiring a reinstall, all based off almost identical code.
[17:44] <persia> If someone has good patches, we should be integrating those.  If it's just look&feel, surely we can come to agreement.
[17:49] <ailo_at> persia: We should identify the reasons to why people would want to choose another distro over US. I'm sure many reasons are pretty trivial. PA, though not crashing so much anymore is still not working with some chips without tweaking. 
[17:51] <ailo_at> There are many things I never do on Linux, so I never see if there are issues. I never us VST plugins for instance (don't know what you need to get those working).
[17:51] <persia> PA is falsely accused in most cases.
[17:52] <ailo_at> I think the most important part for the user is that everything works without tweaking.
[17:53] <persia> But I think that we'd do better to reach out to derivatives, and find out why they are changing things, and whether we could just make a change to solve that problem so they don't need another distro, rather than trying to figure out why users might select something else.
[17:53] <ailo_at> Good point
[17:53] <persia> As a bonus, integrating with the derivatives helps to grow our team.
[17:53] <persia> And that makes it easier to make more changes, and make it better for everyone.
[17:54] <ailo_at> Yes, that would make sense. I think many of the Ubuntu based distros are at least using US kernels if they exist, like Puredyne uses the Karmic rt-kernel. Would make sense to cooperate on some levels.
[17:55] <persia> I strongly suspect that any Ubuntu derivative is using mostly Ubuntu packages.  Some of them have patches, some may have new upstreams, etc.
[17:55] <persia> Most of that work is stuff that we would have done if we had the time/effort/skill
[17:55] <persia> Some of the rest is horrid hacks for which there are better solutions, although more difficult
[17:56] <persia> e.g. fluxbuntu does a lot with dpkg-divert, rather than looking at general solutions, in part because they don't have the time/effort/skill to do it right.
[17:56] <persia> (nothing against fluxbuntu: nice guys, doing interesting things, but limited time makes for less-than-ideal choices sometimes)
[17:57] <persia> And the last bit tends to be branding, themes, etc.
[17:57] <persia> If folk want to do the branding/themes/etc., I think that's fine, and more power to them.
[17:57] <persia> I'd like to have a list of the horrid hacks, so we could try to address them over time (might be a while).
[17:57] <persia> For everything else, I think it ought be done in Ubuntu.
[18:02] <ailo_at> If I've understood correctly realtime will be achived in a secure way in the future (rtkit, right?), and that that will need no further "hacking" in that area (but I suppose all the apps using realtime will need to be updated). Firewire is supposed to work now, thought it won't work with jackd without audio group. What else is there that is a hack now? 
[18:08] <persia> I don't have a complete list of hacks.
[18:08] <persia> What about firewire doesn't work without the "audio" group?
[18:08] <persia> Is that the memlock stuff, or is it broken udev rules?
[18:11] <ailo_at> persia: Firewire will work with jack, but not in realtime, at least that is what I was told from testing. Holstein tried it.
[18:12] <ailo_at> Without audio group
[18:12] <persia> Ah, that's two issues.
[18:12] <persia> 1) jack needs porting to rtkit
[18:13] <persia> 2) jack wants memlock support, and there's no memkit, nor does rtkit support memlock.
[18:13] <persia> Nothing to do with firewire.
[18:13] <ailo_at> Just tried Puredyne i386 version on a AMD64 machine. I get roughly the same performance as with the -lowlatency kernel.
[18:15] <ailo_at> Roughly the same performance as with Natty and the -lowlatency that is
[18:16] <ailo_at> This was using a live USB made with Unetbootin.
[18:16] <persia> OK.  Let's turn that into something actionable.
[18:16] <ailo_at> So, I don't think there aught to be any performance problems, other than that apps may take longer to load.
[18:16] <persia> So, which kernel was that?  If not 2.6.38, what happens with a 2.6.38 kernel?
[18:16] <persia> What is the kernel configuration?  Are there things we should consider adjusting?
[18:17] <persia> How is the live environment constructed?  How does that compare with how Ubuntu does live environments?  Do we believe we can generate a similar experience?
[18:17] <ailo_at> persia: Puredyne, with Karmic -rt kernel. Same as UbuntuStudio Karmic (sorry if I am confusing you)
[18:18] <persia> I'm not confused: I'm only asserting that the results you collected are not actionable in terms of Ubuntu Studio Natty.
[18:18] <persia> So, while interesting, this means they don't translate into a way we can improve things, nor do they necessarily indicate that we should or would expect similar results.
[18:18] <ailo_at> I've used the Puredyne before, and I get the same performance when installing on Hard Disk as well. This was around 2.5ms latency.
[18:24] <ailo_at> I have no idea about how they make the iso. Never done that myself. They use their own script to create a live USB on which you can install programs.
[18:26] <persia> Right.  Needs comparison of that script to the ones Ubuntu uses for the live environment.
[18:27] <persia> There may be differences in how they do things that impact performance.
[18:28] <ailo_at> Here's the script that creates the live device (usb) http://paste.ubuntu.com/568364/
[18:30] <ailo_at> I tried running the script on Maverick, and it didn't work. Either on Karmic or Lucid (could be they updated it since).
[18:30] <ailo_at> They're working on a gui app for it too
[18:31] <ailo_at> If you have bzr installed you can just do "bzr branch lp:bouilloncube" to get the whole thing
[18:33] <persia> The process by which they generate the bootable image isn't the interesting part.  The interesting part is the set of scripts that are used on boot to create the live experience.
[18:38] <ailo_at> persia: Get the ISO maker scripts "bzr branch lp:broth"
[18:40] <persia> I'm in the middle of a few things just now.  Maybe you could look, and compare?
[18:40] <ailo_at> Wheres the Ubuntu script?
[18:43] <persia> livecd-rootfs constructs the liveFS, and casper has the master control scripts that perform the limited installation-like steps to turn that into a live environment (like creating the user "ubuntu", etc.)
[18:55] <ailo_at> Could this be a place to start? https://help.ubuntu.com/community/LiveCDCustomizationFromScratch
[20:21] <scott-work> maybe we should solicit some ideas from users to see what they desire
[20:24] <scott-work> the more i think about this maybe we should ask
[20:24] <scott-work> we have lots of ideas:  live media, updates for -controls, maybe an upgrade pacakge
[20:25] <scott-work> would be nice to know that we are aligned with user's wants
[20:26] <ailo_at> scott-work: Maybe we could ask on lau list. Ask them what they think is most important with a multimedia distro? And what about Graphics and Video, where is the forum for those people?
[20:27] <scott-work> LAU would be a good place to ask for a generalized audience, we can also hit the ubuntustudio-user and ubuntustudio-devel lists as well
[20:27] <scott-work> post in ubuntu forums
[20:28] <scott-work> i've already started a loose questioning in other channels i frequent
[20:28] <scott-work> we might also be able to turn this into a recruiting tool
[20:28] <scott-work> if someone is feeling very strongly about a particular item, we can invite them to help do it :)
[20:28] <scott-work> especially if what they feel strongly about is in the minority and we can't devote enough people to it
[20:29] <ailo_at> Sounds like good ideas to me
[20:29] <scott-work> my first reaction is to list some of the ideas we have, then to ask people what the  most pressing idea is to them, even if it isn't on that list
[20:30] <ailo_at> I think asking other distro people, like persia suggested is a good thing too.
[20:30] <scott-work> or should be just ask them cold..."what would you like to see improved in ubuntu studio?"
[20:30] <scott-work> after february and rpm challenge i wanted to distro hop a few to see what they are doing
[20:30] <scott-work> but we could certainly email them and see if we can incorprate what they do
[20:30] <persia> ailo_at, I think they are different exercises.  Asking users is just general feedback.  Talking to derivatives should be focused clearly on recruiting: we need to understand why they aren't part of Ubuntu Studio, and fix that.
[20:30] <scott-work> mcinnis from dream studio has already contacted me about a few things
[20:36] <scott-work> persia: i would like for us to contact other distros and ask them for collaboration, that is a great idea to foster innovation, improvement, and development :)
[20:36] <scott-work> this is analogous as "upstream" asking for patches
[20:52] <scott-work> i'd like to also play with linux mint and crunch bang as "bases" to build a recording setup during march, to just experiement while i am also playing with avlinux, kxstudio, dream studio, et al
[20:53] <persia> I'm not sure you'll encounter much difference for derivatives that don't focus on audio production.
[20:54] <persia> It's probably just our packages.
[20:55] <ailo_at> As I've said before, I would like all these people to collaborate at least on the "under the hood" stuff.
[20:56] <ailo_at> The multimedia based distro people that is
[20:59] <falktx> ailo_at: i'm already starting to contribute
[20:59] <scott-work> ailo_at: i think that most of the derivatives are not really into the "under the hood" stuff except for falk
[20:59] <scott-work> lol, there he is
[20:59] <falktx> ailo_at: scott-work: http://img340.imageshack.us/img340/6603/scr004.jpg
[20:59] <falktx> hehe
[20:59] <falktx> ^ that is my new tool - a multi-plugin host
[21:00]  * falktx stole the knobs from calf... :)
[21:00] <falktx> the LV2 is killing me
[21:02] <ailo_at> Looks nice. Will vst plugs work too?
[21:04] <falktx> ailo_at: yes, windows plugins too
[21:04] <falktx> native and windows
[21:04] <falktx> ailo_at: I also want to support LMMS plugins
[21:04] <falktx> ailo_at: the gui is written in python, and uses (loads) a backend c++ code, making it possible to use many type of plugins (native, windows, native 32bit on 64bit, etc)
[21:05] <ailo_at> Nice to have everything in one place.
[21:32] <scott-work> if we are going to talk to derivatives of ubuntu studio and solicit collaboration, which distros are these?
[21:33] <scott-work> someone mentioned pure-dyne or dyne-bolic?
[21:33] <scott-work> tango studio is one i belive
[21:33] <scott-work> av linux is another i believe
[21:33] <scott-work> and dream studio
[21:36] <ailo_at> scott-work: With KXStudio, those are the only ones I know of. Puredyne has their own irc channel and mail list. Both are low traffic, but the guys in charge I think are on them both.
[21:37] <ailo_at> I know PureDyne do some kernel hacking, but they chose to use the Karmic -rt kernel for their last release.