Discussion:
FC2 release dates
seth vidal
2003-11-06 15:42:10 UTC
Permalink
Hi,
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.

If we're looking at 4-6 months how about:

Betas around Valentine's Day and St Patrick's Day:
Freeze for the second week of April
Maybe try for a release on Tax Day?


How does that sound.

Very holiday-centric.

-sv
Chris Adams
2003-11-06 16:04:31 UTC
Permalink
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
--
Chris Adams <***@hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.
Dan Young
2003-11-06 16:15:02 UTC
Permalink
Post by Chris Adams
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
See http://fedora.redhat.com/participate/schedule/

2.6 is mentioned for FC2, but ought not keep it from shipping if it
isn't ready in time. Part of the philosophy for Fedora is regular
time-based releases. See the roadmap, about middle of the page:

http://fedora.redhat.com/participate/roadmap/

-Dan Young
-Parkrose School District
seth vidal
2003-11-06 16:18:22 UTC
Permalink
Post by Chris Adams
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
Well it looks like gnome 2.6 will be in late betas in feb/march
http://www.gnome.org/start/2.5/

Might be a good idea to see where that will be and how badly it break
compat with 2.4/2.2.

So it's looking very 2.6-y

kde 3.2 should be in stable for about 3 months by the beta - so that
would be good.

What's mozilla's roadmap look like?

Any other major programs that should be kept in mind?
-sv
Julien Olivier
2003-11-06 17:16:47 UTC
Permalink
Post by seth vidal
Well it looks like gnome 2.6 will be in late betas in feb/march
http://www.gnome.org/start/2.5/
Might be a good idea to see where that will be and how badly it break
compat with 2.4/2.2.
So it's looking very 2.6-y
kde 3.2 should be in stable for about 3 months by the beta - so that
would be good.
What's mozilla's roadmap look like?
Any other major programs that should be kept in mind?
What about OpenOffice ? Any chance to have a Ximian-ised OpenOffice in
Fedora Core 2 ?
Post by seth vidal
-sv
--
fedora-devel-list mailing list
http://www.redhat.com/mailman/listinfo/fedora-devel-list
--
Julien Olivier <***@altern.org>
seth vidal
2003-11-06 17:58:49 UTC
Permalink
Post by Julien Olivier
What about OpenOffice ? Any chance to have a Ximian-ised OpenOffice in
Fedora Core 2 ?
I think Dan Williams is currently Captain OpenOffice at RH - might be
worth getting his input.

-sv
Jeremy Katz
2003-11-07 00:03:48 UTC
Permalink
Post by Julien Olivier
What about OpenOffice ? Any chance to have a Ximian-ised OpenOffice in
Fedora Core 2 ?
Dan Williams is working closely with Michael Meeks from Ximian to get
the things which make sense integrated into the Fedora packages.

Cheers,

Jeremy
Chuck Wolber
2003-11-07 00:18:58 UTC
Permalink
Post by Jeremy Katz
Post by Julien Olivier
What about OpenOffice ? Any chance to have a Ximian-ised OpenOffice in
Fedora Core 2 ?
Dan Williams is working closely with Michael Meeks from Ximian to get
the things which make sense integrated into the Fedora packages.
Just as an anecdote, I'd be curious to see how much of Ximian gets into
Fedora now that Novell has made it clear that their acquisition of SuSE
puts them in direct competition with RedHat.

Oh yeah, and consider this my vote for including a full load of Ximian in
FC ;)

-Chuck
--
Quantum Linux Laboratories - ACCELERATING Business with Open Technology
* Education | -=^ Ad Astra Per Aspera ^=-
* Integration | http://www.quantumlinux.com
* Support | ***@quantumlinux.com

A: Because we read from top to bottom, left to right.
Q: Why should i start my reply below the quoted text?
Michel Alexandre Salim
2003-11-07 04:30:04 UTC
Permalink
Post by Julien Olivier
What about OpenOffice ? Any chance to have a Ximian-ised OpenOffice in
Fedora Core 2 ?
As long as the default file format is not .doc !

Regards,

Michel
Chris Ricker
2003-11-06 17:18:32 UTC
Permalink
Post by seth vidal
Any other major programs that should be kept in mind?
I'd like to see OpenGroupWare merged, and maybe Clam AV to go with
spamassassin. Having an easy to setup, integrated spam / virus / groupware
solution would popularize Fedora in a lot of ways.

Harald @ RH has RPMs for OpenGroupWare, and I've got a mostly working RPM
for clam w/ Postfix. I suppose I can break down, install sendmail, and make
it work with either....

This might be more something for the Extras branch, but I at least think
it's significant enough to go in Core....

later,
chris
seth vidal
2003-11-06 18:01:41 UTC
Permalink
Post by Chris Ricker
Post by seth vidal
Any other major programs that should be kept in mind?
I'd like to see OpenGroupWare merged, and maybe Clam AV to go with
spamassassin. Having an easy to setup, integrated spam / virus / groupware
solution would popularize Fedora in a lot of ways.
for clam w/ Postfix. I suppose I can break down, install sendmail, and make
it work with either....
This might be more something for the Extras branch, but I at least think
it's significant enough to go in Core....
I'd like to reiterate jef spaleta's point and the point of fedora's
release cycle.

TIME BASED. The point is to get something out there and people using it
so it can be tested.

So april 14th (the day before US tax day) is just about 6 months out. It
means being able to get gnome in under the wire and should provide a
good opportunity for a more baked 2.6 kernel.

I think everything else should be targeted at stable fedora extras.

-sv
Vincent
2003-11-07 01:20:48 UTC
Permalink
On Thu, 06 Nov 2003 11:18:22 -0500
Post by seth vidal
Post by Chris Adams
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
Well it looks like gnome 2.6 will be in late betas in feb/march
http://www.gnome.org/start/2.5/
Might be a good idea to see where that will be and how badly it break
compat with 2.4/2.2.
So it's looking very 2.6-y
kde 3.2 should be in stable for about 3 months by the beta - so that
would be good.
What's mozilla's roadmap look like?
Any other major programs that should be kept in mind?
-sv
I think April is perfect. the EOL of RedHat 9 will be there and if these ppl switch
Fedora is where we'd want them to go right? kernel 2.6, Gnome 2.6, etc plus a long
6 month release gives kernel2.6 time to iron out some things, this should _not_ be a quick
release. It is probably going to be a defining release for Fedora most ppl are waiting
for a second release to see where its really going to stand in the market.
David Kewley
2003-11-07 01:55:25 UTC
Permalink
Post by Vincent
I think April is perfect. the EOL of RedHat 9 will be there and if these
ppl switch Fedora is where we'd want them to go right? kernel 2.6, Gnome
2.6, etc plus a long 6 month release gives kernel2.6 time to iron out some
things, this should _not_ be a quick release. It is probably going to be a
defining release for Fedora most ppl are waiting for a second release to
see where its really going to stand in the market.
However, people in larger or more conservative environments will need to do
some heavy testing of FC2 before they roll it out, and that will take longer
than two weeks (e.g. the two weeks between Tax Day and the RHL9 EOL).

So those folks, who are already in a bind with "what version do I standardize
on" will be basically *forced* to go with FC1 as an intermediary (since its
EOL will be later than RHL9's), or else use a different distro altogether,
which means that FC loses a "customer".

I'm in a not-too-conservative environment. But my colleagues and I will
already need to update to RHL9 or FC1 before the end of December, then to FC1
before the end of April. Or else depend on community-supplied security
updates, which still looks like a very shaky proposition (I hope it will
stabilize, but it's too soon to tell).

Many folks are going to have to change their local release policies and
methodology. Our choices seem to be:

* roll out new releases *far* more often than we're used to, based on FC
* fork over some significant cash e.g. for RHEL licenses (which many can't
afford)
* try to make our own distro based on the RHEL srpm's (which feels like
cheating RH, but if it works, it could be a practical solution for us)
* rely on community security updates (so far, too little concreteness to make
an organization's plans around this!)
* leave RH and Fedora

This is a *big* decision for a *lot* of folks.

So far my colleagues and I are still planning to stick with RHL (until EOL)
and FC, but as I say, I anticipate we'll need to majorly rework our upgrade
methods since we're going to have to do them more often, and we cannot spend
a majority of our time adapting to the latest-greatest.

Sorry this morphed from a discussion of FC2 release dates to a discussion of
how to manage a Linux shop in this new age of flux. But they *are* tied
together. If the FC2 release can be planned to allow shops a somewhat
reasonable roll-out plan (say, April 1), that'd be good.

I do not at all intend this to be a complaint or whatever, just pointing out
of the reality that lots of folks are facing now. Anything we can do to
encourage folks to stay with RH and FC is a good thing.

So long as RH doesn't have to spend cash on it. ;)

David
seth vidal
2003-11-08 01:33:40 UTC
Permalink
Post by David Kewley
So those folks, who are already in a bind with "what version do I standardize
on" will be basically *forced* to go with FC1 as an intermediary (since its
EOL will be later than RHL9's), or else use a different distro altogether,
which means that FC loses a "customer".
I'm in a not-too-conservative environment. But my colleagues and I will
already need to update to RHL9 or FC1 before the end of December, then to FC1
before the end of April. Or else depend on community-supplied security
updates, which still looks like a very shaky proposition (I hope it will
stabilize, but it's too soon to tell).
And those people should pay close attention to Fedora Legacy and try to
help themselves and others when possible.

-sv
Paul Gear
2003-11-08 05:07:36 UTC
Permalink
Post by David Kewley
...
Many folks are going to have to change their local release policies and
methodology.
It seems a lot of us are in this position, and none of the options are good.
Post by David Kewley
* roll out new releases *far* more often than we're used to, based on FC
Can't do that - not enough hours in the day already.
Post by David Kewley
* fork over some significant cash e.g. for RHEL licenses (which many can't
afford)
Definitely not affordable. Based on the pricing i've seen (here in
.au), Solaris, Mac OS X, and NetWare are all cheaper, but none of them
thrill me as a good alternative to RHL. For an academic institution, i
can get an *unlimited site license* for NetWare for less than the cost
of two RHEL ES licenses. It's possible that even Micro$oft licensing
would work out cheaper in some scenarios.
Post by David Kewley
* try to make our own distro based on the RHEL srpm's (which feels like
cheating RH, but if it works, it could be a practical solution for us)
When they're releasing SRPMS, how can you say you're cheating them? The
process is working as designed.
Post by David Kewley
* rely on community security updates (so far, too little concreteness to make
an organization's plans around this!)
A variant on this would be to roll your own security updates from the
upstream sources and the sources in FC1.
Post by David Kewley
* leave RH and Fedora
Unless the community effort to produce a modified RHEL distribution is
very successful, i see this as being the best of a bunch of bad
alternatives. I'll be looking at Mandrake and possibly SuSE over the
Dec-Jan period. If i switch, i'll be switching my home, work, and
another .edu where i volunteer over to the new distribution - i simply
can't afford the time committment to
Post by David Kewley
This is a *big* decision for a *lot* of folks.
Is there a way we could get together and try to make some co-ordinated
effort to help RH understand what they've done in alienating the
.edu/.org sector?

And of course, if that fails, making a co-ordinated effort to help each
other by providing a more stable FC or a free RHEL derivative would be a
reasonable alternative.

BTW, has there been any discussion on backporting RHEL features into FC?
Does RH have a policy on this? It seems to me that getting FC to a
state where security errata from RHEL could be applied to it would be a
reasonable compromise.
--
Paul
http://paulgear.webhop.net
--
A: Because we read from top to bottom, left to right.
Q: Why should i start my email reply *below* the quoted text?
Bill Anderson
2003-11-07 22:56:34 UTC
Permalink
Post by seth vidal
Any other major programs that should be kept in mind?
An installer supporting XFS and filesystems other than ext2/3 during
install.
--
Bill Anderson
RHCE #807302597505773
***@noreboots.com
Jesse Keating
2003-11-07 23:07:42 UTC
Permalink
Post by Bill Anderson
An installer supporting XFS and filesystems other than ext2/3 during
install.
ReiserFS and JFS are already options (linux reiserfs jfs). This time RH
even included mkfs.jfs in the installer root file system (;

XFS is the next biggy though.
--
Jesse Keating RHCE MCSE (geek.j2solutions.net)
Fedora Legacy Team (www.fedora.us/wiki/FedoraLegacy)
Mondo DevTeam (www.mondorescue.org)
GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub)

Was I helpful? Let others know:
http://svcs.affero.net/rm.php?r=jkeating
Jeremy Katz
2003-11-07 23:13:49 UTC
Permalink
Post by Bill Anderson
Post by seth vidal
Any other major programs that should be kept in mind?
An installer supporting XFS and filesystems other than ext2/3 during
install.
The installer pretty much supports this all already, mostly just depends
on kernel bits being there (okay, there are a few xfs pieces I haven't
added mostly because of a lack of ability to test them since the kernel
doesn't have support :-)

Jeremy
Hugo Cisneiros
2003-11-06 19:52:25 UTC
Permalink
Post by Chris Adams
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
I think a good choice is to provide native ACL support for the file
system. many people I know want this feature a lot to control their
files on server, specially when used by large groups of people.

I don't know if the implementation is stable enough, but I think when
kernel 2.6 arrives, it'll be a good "ride" to implement native ACL
support in the file system.

I mean native = easy access and documented procedure on how to work with
ACL on files and such :)

[]'s
Hugo
André Kelpe
2003-11-06 20:23:53 UTC
Permalink
Post by Hugo Cisneiros
I think a good choice is to provide native ACL support for the file
system. many people I know want this feature a lot to control their
files on server, specially when used by large groups of people.
I don't know if the implementation is stable enough, but I think when
kernel 2.6 arrives, it'll be a good "ride" to implement native ACL
support in the file system.
I mean native = easy access and documented procedure on how to work with
ACL on files and such :)
Full ACK! ACL is something really important, because it is impossible to
"emulate" ACL with normal user/group permission. IMO this should be one
of the first points on the TODO-list.

André
Michel Alexandre Salim
2003-11-07 04:28:40 UTC
Permalink
Post by Hugo Cisneiros
I don't know if the implementation is stable enough, but I think when
kernel 2.6 arrives, it'll be a good "ride" to implement native ACL
support in the file system.
I mean native = easy access and documented procedure on how to work with
ACL on files and such :)
We should ideally be able to select files in Nautilus, right-click,
select Properties, select Permissions, then get an NT/2k-style tool to
add/remove ACLs. Casual users would then be able to easily share
selected files with others instead of having to mess with setfacl (or,
as needed now, ask an admin to create a new group everytime files need
to be shared. Ugh)

If ACL is a stated goal for FC2 I'll bugzilla this request.

Regards,

Michel
Dave Jones
2003-11-09 22:32:05 UTC
Permalink
Post by Michel Alexandre Salim
Post by Hugo Cisneiros
I mean native = easy access and documented procedure on how to work with
ACL on files and such :)
We should ideally be able to select files in Nautilus, right-click,
select Properties, select Permissions, then get an NT/2k-style tool to
add/remove ACLs. Casual users would then be able to easily share
selected files with others instead of having to mess with setfacl (or,
as needed now, ask an admin to create a new group everytime files need
to be shared. Ugh)
If ACL is a stated goal for FC2 I'll bugzilla this request.
Go for it.
The 2.6 kernel features ACL's without us having to add extra patches,
so this is something that's incredibly likely to show up in FC2.

Dave
Jef Spaleta
2003-11-06 16:25:32 UTC
Permalink
Post by Chris Adams
Hmm, I think an idea of what the goal of FC2 will be should come first.
An obvious one: switch to kernel 2.6? The goals will somewhat drive the
dates.
You development people always look at things in such a myopic way, like
its the content that actually matters as compared to the perception. Yer
totally missing the obvious PR opportunities here.

What better way to say i love you...than with a valentines fedora beta.
Kiss me I'm Irish AND i run Fedora beta2.

And the big PR winner. A release just before the USA tax-day...would
mean you could get LUGs to burn cds and hand them out to the people
standing in line at the big post-offices trying to get their taxes in
under the wire. Can't you see the obvious PR spin involving the concept
of unfair taxation and fedora linux as a generously offered solution to
that emotional and economic heartache.

Successful volunteer movements typically have an eye out for useful PR
opportunities. Maybe its not a driving force...but its a mitigating
one....and something to think about.


-jef"the 5% nation of Casio tones"spaleta
Chris Ricker
2003-11-06 17:29:21 UTC
Permalink
Post by Jef Spaleta
Successful volunteer movements typically have an eye out for useful PR
opportunities. Maybe its not a driving force...but its a mitigating
one....and something to think about.
I think if marketing were a driving concern for this project, a more
marketable name for releases than "Fedora Core" would have been chosen. I
have a classful of students here who just watched me install it. Comments,
in order, on watching it boot for the first time:

0. Graphical bootup was universally liked
1. Click-through licensing in firstboot was either amusing or annoying,
depending on the student
2. Universal confusion as to why it said Fedora Core all over the place.
Even after I explained the Core / Extras / Legacy / etc. business, still
disliking.
Post by Jef Spaleta
-jef"the 5% nation of Casio tones"spaleta
You should've gone for "the 5% nation of harmful free radicals" instead ;-)

later,
chris
seth vidal
2003-11-06 18:44:34 UTC
Permalink
Post by Jef Spaleta
What better way to say i love you...than with a valentines fedora beta.
Kiss me I'm Irish AND i run Fedora beta2.
This is obviously such a wonderful tag line for a beta announcement it
is painful.

+1
Post by Jef Spaleta
And the big PR winner. A release just before the USA tax-day...would
mean you could get LUGs to burn cds and hand them out to the people
standing in line at the big post-offices trying to get their taxes in
under the wire. Can't you see the obvious PR spin involving the concept
of unfair taxation and fedora linux as a generously offered solution to
that emotional and economic heartache.
Gotta make sure gnucash is working though. ;)
Post by Jef Spaleta
Successful volunteer movements typically have an eye out for useful PR
opportunities. Maybe its not a driving force...but its a mitigating
one....and something to think about.
Take another +1 for Gryffindor Mr Spaleta.

-sv
Bill Nottingham
2003-11-06 19:05:02 UTC
Permalink
Post by seth vidal
Gotta make sure gnucash is working though. ;)
Uh-oh, then we have to wait for gnome2 gnucash.

Bill
Jef Spaleta
2003-11-06 18:01:55 UTC
Permalink
Post by Chris Ricker
I think if marketing were a driving concern for this project, a more
marketable name for releases than "Fedora Core" would have been chosen. I
have a classful of students here who just watched me install it. Comments,
Now, I could probably argue that there is quite a lot of room in this process to implement
'a lessons learned' phase where we sit back and constructively try to navel gaze about how
certain decisions were good or bad or worth reconsidering. But I'm not
going to argue for anything like that till at least FC3. And I'm not
going to even try to imply that 'Core' is bad till we see 'Extras /
Alternatives' media out in the wild to complicate things.

I'd rather not fight battles over decisions already made just yet...when
there are so many important decisions yet to make. I'm going to
concentrate on making sure the remaining decisions are as multi-facted,
win-win, opportunity advantageous as possible.


-jef"someone didn't catch my song quote reference..oh well"spaleta
Chris Ricker
2003-11-07 13:02:09 UTC
Permalink
Post by Jef Spaleta
Post by Chris Ricker
I think if marketing were a driving concern for this project, a more
marketable name for releases than "Fedora Core" would have been chosen. I
have a classful of students here who just watched me install it. Comments,
Now, I could probably argue that there is quite a lot of room in this process to implement
'a lessons learned' phase where we sit back and constructively try to navel gaze about how
certain decisions were good or bad or worth reconsidering. But I'm not
going to argue for anything like that till at least FC3. And I'm not
going to even try to imply that 'Core' is bad till we see 'Extras /
Alternatives' media out in the wild to complicate things.
I'd rather not fight battles over decisions already made just yet...when
there are so many important decisions yet to make. I'm going to
concentrate on making sure the remaining decisions are as multi-facted,
win-win, opportunity advantageous as possible.
You're missing my point. Criteria which should be used for decisions are the
ones which are important to the project. I don't see marketability as being
an important criteria here. The naming is just one evidence of that....

As such, release dates should be based on real criteria which actually
matter to the project, not on unimportant factors like ease of marketing
tie-ins.
Post by Jef Spaleta
-jef"someone didn't catch my song quote reference..oh well"spaleta
I got it. It wasn't applicable.

later,
chris
Jeremy Katz
2003-11-07 00:13:25 UTC
Permalink
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
I tend to agree, although they're test releases, not betas :-)
Post by seth vidal
Freeze for the second week of April
Maybe try for a release on Tax Day?
How does that sound.
I would personally prefer to get 2.6 out sooner rather than later. To
throw out something as an alternative plan for a faster release,
something like the following for availability of milestones could work.
It's definitely a more aggressive suggestion :)

Test1 - January 13
Test2 - February 3
Test3/RC - March 1
GA - March 22

Then, extrapolate back roughly a week earlier for freeze dates. I'd
love to actually get earlier, but that would require either less test
releases (which strikes me as a bad idea for 2.6), less time between
them (same) or getting test 1 out before Christmas (likely to be really
really hard to pull off)

Anyway, my $0.02 on the matter

Jeremy
seth vidal
2003-11-07 00:19:52 UTC
Permalink
Post by Jeremy Katz
I tend to agree, although they're test releases, not betas :-)
what's the distinction in nomenclature?
Post by Jeremy Katz
I would personally prefer to get 2.6 out sooner rather than later. To
throw out something as an alternative plan for a faster release,
something like the following for availability of milestones could work.
It's definitely a more aggressive suggestion :)
Test1 - January 13
Test2 - February 3
Test3/RC - March 1
GA - March 22
I liked the month between test releases better than 2 weeks. If only b/c
the installer(as you well know) will be less tested if there is less
time to sync test isos to do full reinstalls. And for a lot of people it
still takes about a day or 2 to get 3 isos.

And I think the installer+2.6+lvm/evms/etc will be more than enough pain
to iron out.
Post by Jeremy Katz
Then, extrapolate back roughly a week earlier for freeze dates. I'd
love to actually get earlier, but that would require either less test
releases (which strikes me as a bad idea for 2.6), less time between
them (same) or getting test 1 out before Christmas (likely to be really
really hard to pull off)
What's the motive to shorten up to end of March instead of Mid-April?
You'll start get bugs rolling in from the test releases for 2.6. I don't
think there will be any shortage. :)

and the dates you mention don't have humorous holiday jokes involved -
you've got to keep the important stuff in perspective! :)

-sv
Jeremy Katz
2003-11-07 00:34:17 UTC
Permalink
Post by seth vidal
Post by Jeremy Katz
I tend to agree, although they're test releases, not betas :-)
what's the distinction in nomenclature?
It's on http://fedora.redhat.com/participate/terminology.html --
To distinguish Red Hat Enterprise Linux's formal beta program from the
community-based testing of Fedora Core, we will use the term "test"
release instead of "beta" release to refer to test releases.
Post by seth vidal
Post by Jeremy Katz
I would personally prefer to get 2.6 out sooner rather than later. To
throw out something as an alternative plan for a faster release,
something like the following for availability of milestones could work.
It's definitely a more aggressive suggestion :)
[snip]
Post by seth vidal
I liked the month between test releases better than 2 weeks. If only b/c
the installer(as you well know) will be less tested if there is less
time to sync test isos to do full reinstalls. And for a lot of people it
still takes about a day or 2 to get 3 isos.
That's four weeks between all but test3 and final which is a slightly
shorter one. Hopefully, having rawhide continually installable will
help get installer feedback sooner and allow people to test fixes much
sooner. Also, being able to do graphical installs via FTP or HTTP
should be a nice help here as well.
Post by seth vidal
And I think the installer+2.6+lvm/evms/etc will be more than enough pain
to iron out.
Oh, believe me... I have a pretty good idea of how much pain is involved
:/
Post by seth vidal
Post by Jeremy Katz
Then, extrapolate back roughly a week earlier for freeze dates. I'd
love to actually get earlier, but that would require either less test
releases (which strikes me as a bad idea for 2.6), less time between
them (same) or getting test 1 out before Christmas (likely to be really
really hard to pull off)
What's the motive to shorten up to end of March instead of Mid-April?
You'll start get bugs rolling in from the test releases for 2.6. I don't
think there will be any shortage. :)
Because people are already asking for 2.6 stuff now and test releases
aren't the same as an actual release. That's basically the fastest
timeline I can come up with that I think can be hit and still have
something that's robust.

Jeremy
seth vidal
2003-11-07 00:47:52 UTC
Permalink
Post by Jeremy Katz
That's four weeks between all but test3 and final which is a slightly
shorter one. Hopefully, having rawhide continually installable will
help get installer feedback sooner and allow people to test fixes much
sooner. Also, being able to do graphical installs via FTP or HTTP
should be a nice help here as well.
umm continually installable rawhide. hehe.

/me holds his breath. :)
Post by Jeremy Katz
Because people are already asking for 2.6 stuff now and test releases
aren't the same as an actual release. That's basically the fastest
timeline I can come up with that I think can be hit and still have
something that's robust.
hmm Makes a crunch for gnome 2.6 then. Looks like it will just miss it
on that schedule. Maybe I'll talk to luis and see if there are plans for
a 2.4.X release to fix misc bugs.

-sv
Jens Knutson
2003-11-07 01:05:42 UTC
Permalink
Post by seth vidal
Post by Jeremy Katz
Because people are already asking for 2.6 stuff now and test releases
aren't the same as an actual release. That's basically the fastest
timeline I can come up with that I think can be hit and still have
something that's robust.
hmm Makes a crunch for gnome 2.6 then. Looks like it will just miss it
on that schedule. Maybe I'll talk to luis and see if there are plans for
a 2.4.X release to fix misc bugs.
Indeed. Certainly, waiting 2-4 weeks on FC 2 to make sure GNOME 2.6
makes it in, too, would be worth it!

- jck
--
"To turn $100 into $110 is work. To turn $100 million into $110 million
is inevitable"
-- Edgar Bronfman
seth vidal
2003-11-07 01:03:33 UTC
Permalink
Post by seth vidal
hmm Makes a crunch for gnome 2.6 then. Looks like it will just miss it
on that schedule. Maybe I'll talk to luis and see if there are plans for
a 2.4.X release to fix misc bugs.
This is one thing I think might want to be emphasized. I think gnome 2.6
is looking good for being 1. on time and 2. pretty solid - according to
Luis.

Maybe the novell-ximian thing brought Luis some more resources to beat
on bugs, who knows. :)


Worth asking: What group makes this final decision?

-sv
Michael K. Johnson
2003-11-07 20:38:44 UTC
Permalink
Post by seth vidal
Worth asking: What group makes this final decision?
As technical leader, it's my job to fix the fact that our leadership
page on the fedora.redhat.com site is still a draft. That's something
I'll be working on next week (doesn't mean I'll finish next week, but
it's going to be one of the few items at the top of my list, so I'll
probably get behind on email and be high-latency on IRC) and then we
will have the group that can make the call.

Roughly, we'll have a committee who is responsible for listening to
all the arguments and making a decision. It will have folks from
Red Hat and others on it. Red Hat's needs and community needs will
both be part of the decision-making process.

I should mention here my concept of Fedora leadership. I don't
believe that leadership by appointment works very well, at least
in this context. My intention for Fedora leadership is that
we'll be *recognizing* people who are already acting as the
leaders in the community. Not necessarily the people who have
the most to say all the time (<grin>) but the ones who lead
by action, and whose goals as expressed in their actions are
aligned with our goals, the statement of which is available at
http://fedora.redhat.com/about/objectives.html and which we expect
to refine over time.

I want to be clear that Red Hat will be bringing some clear objectives
to the table for this discussion. I'll say right away that a 2.6
kernel as soon as possible has at least three reasons that we care about:
o We have a LOT of end users urgently requesting it.
o Linus has asked us to do a 2.6-kernel-default distribution
absolutely ASAP, and his request carries significant weight.
o Red Hat has an interest in doing a distribution with a 2.6
kernel from the standpoint of our work on our Enterprise
products.

HTH,

michaelkjohnson

"He that composes himself is wiser than he that composes a book."
Linux Application Development -- Ben Franklin
http://people.redhat.com/johnsonm/lad/
seth vidal
2003-11-07 22:28:34 UTC
Permalink
Post by Michael K. Johnson
I should mention here my concept of Fedora leadership. I don't
believe that leadership by appointment works very well, at least
in this context. My intention for Fedora leadership is that
we'll be *recognizing* people who are already acting as the
leaders in the community. Not necessarily the people who have
the most to say all the time (<grin>) but the ones who lead
by action, and whose goals as expressed in their actions are
aligned with our goals, the statement of which is available at
http://fedora.redhat.com/about/objectives.html and which we expect
to refine over time.
This is good to hear. Better than appointing, you're definitely correct.
I think nominations for people worthy of recognition might be useful.
B/c, as you said, you've been busy and might not have noticed people who
have been having an effect. I have a few people in the back of my head
who deserve high praise for putting up with a fair bit of abuse, many of
them at red hat. :)
Post by Michael K. Johnson
I want to be clear that Red Hat will be bringing some clear objectives
to the table for this discussion. I'll say right away that a 2.6
o We have a LOT of end users urgently requesting it.
o Linus has asked us to do a 2.6-kernel-default distribution
absolutely ASAP, and his request carries significant weight.
o Red Hat has an interest in doing a distribution with a 2.6
kernel from the standpoint of our work on our Enterprise
products.
All of those are pressing points and well understood, my only argument
for delaying a bit is to try and include gnome 2.6. - As Jef Spaleta
mentioned on irc - a cool release name is FC2(.6) ;)

I know a month is a long time in free software but from the avg life
cycle of linux 2.2->2.4->2.6 it won't be a very long in the linux
2.6->2.8 cycle. Getting gnome 2.6 would mean more gnome testing and
that's got to be helpful for RHEL, as well.

my thoughts, worth whatever they're worth.

-sv
Post by Michael K. Johnson
HTH,
michaelkjohnson
"He that composes himself is wiser than he that composes a book."
Linux Application Development -- Ben Franklin
http://people.redhat.com/johnsonm/lad/
--
fedora-devel-list mailing list
http://www.redhat.com/mailman/listinfo/fedora-devel-list
seth vidal
2003-11-07 22:36:33 UTC
Permalink
Post by seth vidal
I know a month is a long time in free software but from the avg life
cycle of linux 2.2->2.4->2.6 it won't be a very long in the linux
2.6->2.8 cycle. Getting gnome 2.6 would mean more gnome testing and
that's got to be helpful for RHEL, as well.
clarification of this.

the 'it' in the first sentence refers to '1 month' not to the life cycle
of 2.6->2.8.

My point was that one month wasn't much time considering that the time
between 2.2 and 2.4 and 2.4 and 2.6 was considerable.


-sv
Dag Wieers
2003-11-07 23:29:55 UTC
Permalink
Post by seth vidal
Post by Michael K. Johnson
I want to be clear that Red Hat will be bringing some clear objectives
to the table for this discussion. I'll say right away that a 2.6
o We have a LOT of end users urgently requesting it.
o Linus has asked us to do a 2.6-kernel-default distribution
absolutely ASAP, and his request carries significant weight.
o Red Hat has an interest in doing a distribution with a 2.6
kernel from the standpoint of our work on our Enterprise
products.
All of those are pressing points and well understood, my only argument
for delaying a bit is to try and include gnome 2.6. - As Jef Spaleta
mentioned on irc - a cool release name is FC2(.6) ;)
I know a month is a long time in free software but from the avg life
cycle of linux 2.2->2.4->2.6 it won't be a very long in the linux
2.6->2.8 cycle. Getting gnome 2.6 would mean more gnome testing and
that's got to be helpful for RHEL, as well.
It would be nice to have at least one release where the kernel 2.6 may be
optional. Arjan's test kernels for RH9 (or rather Rawhide) fits into this
scheme. Enlarging the test-group without sacrificing too many (all?)
people ;)

-- dag wieers, ***@wieers.com, http://dag.wieers.com/ --
[Any errors in spelling, tact or fact are transmission errors]
seth vidal
2003-11-07 23:51:15 UTC
Permalink
Post by Dag Wieers
It would be nice to have at least one release where the kernel 2.6 may be
optional. Arjan's test kernels for RH9 (or rather Rawhide) fits into this
scheme. Enlarging the test-group without sacrificing too many (all?)
people ;)
I think that release is FC1 + the arjan test kernel repository.

-sv
Dag Wieers
2003-11-08 00:20:35 UTC
Permalink
Post by seth vidal
Post by Dag Wieers
It would be nice to have at least one release where the kernel 2.6 may be
optional. Arjan's test kernels for RH9 (or rather Rawhide) fits into this
scheme. Enlarging the test-group without sacrificing too many (all?)
people ;)
I think that release is FC1 + the arjan test kernel repository.
Well, making the 2.6 kernel optional (maybe through apt) without requiring
to add a new repository would help. I would consider adding Arjan's
packages to FC1 as too experimental ('as there must be a reason why Fedora
doesn't add it, right ?')

Adding it to the updates (or testing) would probably reach a bigger target
group. The next release could default to 2.6 and optionally 2.4 (for
hardware or software that worked on 2.4 but doesn't on 2.6 anymore)

It's a small difference (maybe even psychological), but if you want to
test it in advance, the more people/hardware you can test it on, the
better the next release will integrate.

-- dag wieers, ***@wieers.com, http://dag.wieers.com/ --
[Any errors in spelling, tact or fact are transmission errors]
Jeremy Katz
2003-11-08 02:15:50 UTC
Permalink
Post by Dag Wieers
Adding it to the updates (or testing) would probably reach a bigger target
group. The next release could default to 2.6 and optionally 2.4 (for
hardware or software that worked on 2.4 but doesn't on 2.6 anymore)
Optional kernels are a losing proposition. The installer can really
only run with one kernel (otherwise, there's a large number of things
that has to be conditional on kernel version which leads to absurd
amounts of pain, especially as only one path will get tested) at which
point 2.6 *has* to work for people's hardware so that they can install.
Even if you discount that, how do you present an option like that?
[ ] My hardware sucks, use 2.4
[*] Use 2.6 like the rest of the world
You don't even know if your hardware is going to have problems until
you're already using 2.6.

Also, there are some things that would be nice to do along with 2.6 that
will then require the system to be using 2.6 -- things like switching to
ALSA across the board, using udev to get persistent device naming, and
others.
Post by Dag Wieers
It's a small difference (maybe even psychological), but if you want to
test it in advance, the more people/hardware you can test it on, the
better the next release will integrate.
Having things so that both 2.4 and 2.6 work mean that the integration of
2.6 isn't going to be as smooth.

Cheers,

Jeremy
Eric S. Raymond
2003-11-08 02:30:41 UTC
Permalink
Post by Jeremy Katz
Having things so that both 2.4 and 2.6 work mean that the integration of
2.6 isn't going to be as smooth.
Agreed. I don't know if my opinion counts for anything in this, nor
even if it should. But I'd like to see FC2 go for 2.6 all the way
unless 2.6 has showstopper problems.

In general I think Fedora can, and should, take a bit more aggressive
forward position than stock Red Hat releases did. Not that I'm knocking
Red Hat's policy, it made a lot of sense for a packaged consumer product.
But the tradeoffs are a little different now. One important difference is
that apt-get/yum reduces the cost of being too aggressive in a major
release by making fix propagation easier.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
Michael K. Johnson
2003-11-10 21:13:46 UTC
Permalink
Post by Eric S. Raymond
Post by Jeremy Katz
Having things so that both 2.4 and 2.6 work mean that the integration of
2.6 isn't going to be as smooth.
Agreed. I don't know if my opinion counts for anything in this, nor
even if it should. But I'd like to see FC2 go for 2.6 all the way
unless 2.6 has showstopper problems.
I'll stick my neck out and say: It's either 2.4 or 2.6, no middle
ground. If 2.6 isn't good enough to not need an alternative, then
we've got a problem. But I don't think we'll have a problem; 2.6
is progressing much more nicely than 2.4 did from my viewpoint.

michaelkjohnson

"He that composes himself is wiser than he that composes a book."
Linux Application Development -- Ben Franklin
http://people.redhat.com/johnsonm/lad/

Paul Heinlein
2003-11-08 14:48:19 UTC
Permalink
Post by Jeremy Katz
Optional kernels are a losing proposition.
Aye! I've not used 2.6 in any significant way, but recently I
installed a minimal Debian stable on a machine at home, with plans to
immediate report my sources.list at testing trees and do a fuller
install.

Debian stable, of course, still ships with 2.2.old, and the installer
decided to use the rtl8139 modules for my NIC. It worked fine. Then I
upgraded to a 2.4 kernel. Whoops, rtl8139 was nowhere to be found. It
didn't take me too long to locate and load the newer 8139too module,
but I pretty much knew what I was doing.

Had I not, it would have been ugly.

Again, I don't know if that sort of thing is an issue in a 2.4 -> 2.6
migration, since I haven't really done much with 2.6, but my hunch is
that it's just a problem waiting to happen.

--Paul Heinlein <***@madboa.com>
Michael K. Johnson
2003-11-10 20:53:32 UTC
Permalink
I took the weekend off (horrors!) so this may seem a bit late :-)
Post by seth vidal
This is good to hear. Better than appointing, you're definitely correct.
I think nominations for people worthy of recognition might be useful.
B/c, as you said, you've been busy and might not have noticed people who
have been having an effect. I have a few people in the back of my head
who deserve high praise for putting up with a fair bit of abuse, many of
them at red hat. :)
:-)

My first goal is to refine the definitions of the groups. I think that
we'll end up with different names and a slightly different structure;
we might not have the "Advisory Committee" per se (name is confusing
and some of the delineation seems a bit artificial) so I'm making some
proposals that I need to get Steering Committee approval for.

My intent is to roll out better definitions soon before worrying about
final membership.
Post by seth vidal
All of those are pressing points and well understood, my only argument
for delaying a bit is to try and include gnome 2.6. - As Jef Spaleta
mentioned on irc - a cool release name is FC2(.6) ;)
I know a month is a long time in free software but from the avg life
cycle of linux 2.2->2.4->2.6 it won't be a very long in the linux
2.6->2.8 cycle. Getting gnome 2.6 would mean more gnome testing and
that's got to be helpful for RHEL, as well.
my thoughts, worth whatever they're worth.
Absolutely worth consideration. I just wanted to present our reasons
because one of the points of the new regime is to not appear arbitrary
by keeping silent about our reasons... (I don't think we've every
really been arbitrary, but I know it has looked that way sometimes.)

michaelkjohnson

"He that composes himself is wiser than he that composes a book."
Linux Application Development -- Ben Franklin
http://people.redhat.com/johnsonm/lad/
Barry K. Nathan
2003-11-07 09:38:02 UTC
Permalink
Post by seth vidal
hmm Makes a crunch for gnome 2.6 then. Looks like it will just miss it
on that schedule. Maybe I'll talk to luis and see if there are plans for
a 2.4.X release to fix misc bugs.
FWIW, I saw a mail on the Mandrake cooker list asking people to
completely keep GNOME 2.6 stuff out of cooker. It sounds like Mandrake
10.0 may still ship with GNOME 2.4.

Ok, I found the mail here:
http://marc.theaimsgroup.com/?l=mandrake-cooker&m=106811594208408&w=2

(Personally, I don't care too much about how the FC2 schedule goes. FWIW I
was mildly annoyed when FC1 was delayed for GNOME 2.4 -- I wanted to see
a release with RPM 4.2.1 and ExecShield sooner -- but I think it turned
out for the best in the end.)

-Barry K. Nathan <***@pobox.com>
seth vidal
2003-11-07 17:44:37 UTC
Permalink
Post by Barry K. Nathan
FWIW, I saw a mail on the Mandrake cooker list asking people to
completely keep GNOME 2.6 stuff out of cooker. It sounds like Mandrake
10.0 may still ship with GNOME 2.4.
http://marc.theaimsgroup.com/?l=mandrake-cooker&m=106811594208408&w=2
(Personally, I don't care too much about how the FC2 schedule goes. FWIW I
was mildly annoyed when FC1 was delayed for GNOME 2.4 -- I wanted to see
a release with RPM 4.2.1 and ExecShield sooner -- but I think it turned
out for the best in the end.)
Well I guess I'm thinking from two perspectives:

1. it will mean more testing for gnome, which is definitely good.
2. it could mean staying roughly in-sync with gnome which is not bad
3. if fc1 is gnome 2.4 and fc2 is gnome 2.4 then it's a good shot that
fc3 is gnome 2.8 - which is a bigger jump than gnome 2.4->2.6.

Also there is that whole 'things numbered 2.6' which is nice.

-sv
Kjartan Maraas
2003-11-08 13:40:14 UTC
Permalink
Post by seth vidal
Post by Jeremy Katz
That's four weeks between all but test3 and final which is a slightly
shorter one. Hopefully, having rawhide continually installable will
help get installer feedback sooner and allow people to test fixes much
sooner. Also, being able to do graphical installs via FTP or HTTP
should be a nice help here as well.
umm continually installable rawhide. hehe.
/me holds his breath. :)
Post by Jeremy Katz
Because people are already asking for 2.6 stuff now and test releases
aren't the same as an actual release. That's basically the fastest
timeline I can come up with that I think can be hit and still have
something that's robust.
hmm Makes a crunch for gnome 2.6 then. Looks like it will just miss it
on that schedule. Maybe I'll talk to luis and see if there are plans for
a 2.4.X release to fix misc bugs.
There is a 2.4.1 release out or staging at the moment, and I think I'll
get a 2.4.2 release out too even, if there's a sufficient amount of
bugfixes in the near future.

Cheers
Kjartan
The stable GNOME release team :)
seth vidal
2003-11-08 15:50:00 UTC
Permalink
Post by Kjartan Maraas
There is a 2.4.1 release out or staging at the moment, and I think I'll
get a 2.4.2 release out too even, if there's a sufficient amount of
bugfixes in the near future.
what things have been fixed, and how do you think it compares to the 2.6
devel at the moment?

what things are slated to be in 2.6 over and above 2.4?
-sv
Keith Sharp
2003-11-09 11:09:25 UTC
Permalink
Post by seth vidal
Post by Kjartan Maraas
There is a 2.4.1 release out or staging at the moment, and I think I'll
get a 2.4.2 release out too even, if there's a sufficient amount of
bugfixes in the near future.
what things have been fixed, and how do you think it compares to the 2.6
devel at the moment?
what things are slated to be in 2.6 over and above 2.4?
I would guess the Slashdot headline grabbing feature will be GTK+ 2.4
with the new File Chooser. For more detail you'll need to wait until
the call for modules is completed.

Keith.
Gene C.
2003-11-07 08:38:21 UTC
Permalink
Post by Jeremy Katz
Post by seth vidal
I think we should get an estimate on FC2 beta/freeze/release dates
ironed out right now.
I tend to agree, although they're test releases, not betas :-)
Post by seth vidal
Freeze for the second week of April
Maybe try for a release on Tax Day?
How does that sound.
I would personally prefer to get 2.6 out sooner rather than later. To
throw out something as an alternative plan for a faster release,
something like the following for availability of milestones could work.
It's definitely a more aggressive suggestion :)
Test1 - January 13
Test2 - February 3
Test3/RC - March 1
GA - March 22
Then, extrapolate back roughly a week earlier for freeze dates. I'd
love to actually get earlier, but that would require either less test
releases (which strikes me as a bad idea for 2.6), less time between
them (same) or getting test 1 out before Christmas (likely to be really
really hard to pull off)
If the focus of the next round is the 2.6 kernel, then you can get lots of
updates kernel tests via up2date. Are not the test1/test2/test3 releases
really focused on the install process? With the release plus up2date
process, individual packages (including the kernel) get a lot more rounds of
testing than is implied by test release dates.
--
Gene
Loading...