[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [fab] looking at our surrent state a bit

Rahul Sundaram schrieb:
> Thorsten Leemhuis wrote:
>>  * it's not that much present -- we know it exists, but that's often all.
>>  * seems to meet quite seldom and it's hard to see what it does or if
>> there even is progress somewhere
> We havent had a meeting in the last few weeks due to the release work 
> and other things but whenever there is one, the agenda is posted here 
> and post meeting results are available in the wiki and send to 
> fedora-announce list.
> http://fedoraproject.org/wiki/Board/Meetings/
> What else could be done?

IMHO: Meet more often. Get more involved into decisions. Show presence.

>>  * some Extras contributors mentioned to me that the hierarchy in the
>> whole project is not documented properly (does FESCo get orders from the
>> board? Where is the Packaging Committee located in the whole picture?
>> Stuff like that...).
> That's essentially pretty simple but that could be documented better. I 
> will do that. [...]


>>  * why doesn't the board at least now and then meet on irc so other
>> interested parties can watch or comment?
> I wouldnt mind and I heard opinions that phone conversations move much 
> faster. [...]

Well, phone meetings are okay. That's why I suggested "now and then meet
on irc" -- then everyone can get in contact with the board now and then
while it can still get work done. That gets rid of the "secret cabal"
look that some outsiders might have.

>>  But on the other hand it seems to me that the progress in our
>> distribution specific stack (anaconda, config tools, initscripts) is
>> quite slow. 
> Is Anaconda really in the list of slow moving projects? [...]

Yes and no -- The team does a good job, but looking back it took quite
some time until anaconda was capable of using Extras during install. To
long IMHO. Installing packages from CD/DVD also is still lacking.

>> And not only that, also the infrastructure of Fedora for the
>> community (new VCS, let community help in Core, ...) seems to go forward
>> quite slowly (e.g. nearly nothing).
> Jesse Keating is working on setting up a mercurial repository.

I know, and Jesse, Jeremy, Bill and the others are doing great work, but
their days are only 24 hours long, too (and their work days are limited,

>>  The Live-CD is a good example for the problems -- how long are we
>> working on it now without a real result? Much to long!
> Fedora Unity produced some Live CD's which can be considered real 
> results.


> Official CD releases are unfortunately taking a longer time but 
>   if various sub projects would require Red Hat developers to work on 
> them  that would essential mean we would have to prioritize the work.

Exactly that's what I'd like to see.

>>  I also like Fedora Core due to the "Open-Source only" and "Upsteam
>> please" attitude. But most of the normal users only see the
>> disadvantages (nearly no drivers/features that are not upstream in out
>> packages, no ACPI-DSTD in initrd [see also
>> http://hughsient.livejournal.com/5889.html -- that blog entry is a good
>> general example IMHO], no acrobat, no jre from sun, no proprietary
>> drivers from ati/nvidia and not even the firmware for ipw2[12]00 ) that
>> behavior creates -- and at the same time we are AFAICS quite bad when if
>> comes to communicate the "But we are the good guys and that's the
>> disadvantage we have to for being the good guys" to out users (that
>> might give us some bonus points here and there).
> What could we do about that?

Not sure, I'm not a marketing guy. But we need to communicate better
that we're nearly as free as Debian (here and there we are worse, in
other areas we are better afaics). Most people don't know that afaics.

>>  We also don't get a unique "Fedora look and feel" to the world. 
> In the last couple of releases, the logo and the work done in the Fedora 
> artwork team is very well recognized as unique and appealing in many 
> places.

The artwork is great and I really like it, but that's not what I meant.
;-) It was more and introduction to this para:

> "Fedora
>> is about the rapid progress of Free and Open Source software and
>> content." (quote from http://fedoraproject.org/wiki/Objectives ). Well,
>> that's true in some parts of Fedora (nearly always latest KDE, major
>> kernel updates, Gnome Updates to 2.x.0 to 2.x.[0-9], lot's of updates in
>> Extras-Land), but fail in other areas (no gutenprint in FC6 [a lot of
>> printers are not supported due to that], only Firefox 1.5[Ubuntu 6.10
>> shipped two days after FC6 and has Firefox 2.0 and gutenprint] and no
>> sign of a update in Core to FF 2.0, no X.org-Update to 7.1 [even after
>> the proprietary drivers where able to handle it; owners of G965 hardware
>> were left out in the cold without Support in Fedora due to this as the
>> driver for that popular hardware depends on/is shipped in Xorg 7.1],
>> sometimes users have to wait ages to get the latest Gnome (remember FC4)
>> because that's not updated and out schedule isn't aligned to the gnome
>> schedule [in other words: users of Ubuntu get the hard work from a lot
>> of Red-Hat-Gnome-hackers earlier then Fedora Core users --
>> arrggghhhhh]). 
> That's the essence of free software.

I know that, but it seems you didn't understand what I was up to.

>> I especially dislike the behavior for
>>  * Gnome and Firefox as a lot of users are interested to run the latest
>> version of those packages (sure, that's often stupid, but that's how it is)
>>  * X.org and gutenprint, as hardware support is crucial -- that sucks
>> even more as out hardware support in other areas of Fedora is quite good
>> as kernel and packages like sane get updates to new upstream version
>> regularly
> Gutenberg, Xorg 7.1, Firefox 2.0 updates were discussed in fedora-devel 
> list in detail. So I wont rehash that now.
> You want to push all major updates like GNOME

No, I think we should align out schedule to Gnome, as it is a crucial
part of our product.

> and Xorg releases

Xorg: yes. The updates improve hardware support and are often needed to
get the latest Hardware running. And that's why I think why we should
ship them often (still needs to be decided on a case by case basis).

> into 
> updates in general release which is not really feasible if you want some 
> form of stability. Rapid progress does not mean we can push everything 
> into updates.

Sure. That's not what I proposed. But if there are important things
missing (FF 2.0 in FC6; AIGLX in FC5, Gnome Update in FC4) then try to
get a solution that makes installing those software possible easily (the
AIGLX on FC5 was more a disaster because it was poorly maintained).

>>  We still have no "Fedora Core steering Commitee" (see also
>> https://www.redhat.com/archives/fedora-advisory-board/2006-September/msg00079.html
> http://fedoraproject.org/wiki/Core/SteeringCommittee.

I had something like FESCo im mind. E.g. more members (maybe even from
outside of RH), public meetings, summaries to the list.

>> ) -- what core does or how decisions are made it completely in the dark
>> for the Community and that really sucks.
> More public regular meetings might help. I have suggested that.


>> == Fedora Extras ==
>>  * Developers from Core talk to Extras contributors more often these
>> days; still far from prefect, but it's getting better
>>  * the Fedora Directory Server is still not in Core or Extras afaik
> Individual pieces required for FDS like svrcore-devel and mozldap this 
> is already under review in Fedora Extras and the directory server team 
> is working on fixing various aspects like following FHS better, 
> autotools, static libs etc. It wouldnt pass through review without 
> making the developer changes and testing them.

Yeah, warren mentioned something during yesterdays meeting. But it took
quite some time and that's the biggest problem in Fedora -- everything
takes a long time.

>>  * we can't do anything we'd like to do; I hope we can get a bit more
>> support from RH in the future
> What does this mean?

Well, take Comaintainership as example. We need

- acl support in the build system
- a package database to get per dist maintainers
- a better framework to make it easier to watch what the people you
sponsored changes in cvs

That's a lot of work (mainly on the infrastructure side) and I more and
more get the impression that with our current manpower it will take one
or two years until we might have it in place. A little bit help from
someone that getting payed for his work might help

>> == MISC ==
>>  * I got the impression (and LWN readers, too ["hello corbert! "]) that
>> Fedora Legacy is not able to do it's job properly. Maybe it's time to
>> just revamp the whole project?
> How?

Give it a fresh start, a new name (because the Term "Fedora Legacy" has
such a bad fame now), maybe try to get the load reduced (only support
releases with odd number for a longer time, drop old releases).


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]