Re: [gdm-list] update

[Brian Cameron <Brian Cameron Sun COM>]

Jon:

> "a list of requirements that we need GDM to honor"
>
> I'll assume you aren't joking.  You expect an upstream free software
> project to be designed around a list of requirements developed in
> private for a proprietary and closed source product?

Bob was saying that there is a list of requirements that the Sun
Ray team needs GDM to support in order for GDM to work with Sun
Ray.  If these requirements cannot be met, then it seems obvious
that the Sun Ray team will end up using a different display manager
moving forward.

At any rate, I don't think Bob's words were particularly provocative.
It seems poor form to close the door on discussion and requirements
gathering just because Sun Ray is not (yet) an open source project.
Jon, you already said that you anticipate working on supporting related
functionality anyway.  It seems that being informed of Sun Ray
requirements would be a useful thing to keep in mind as you go.

I understand that the stresses of preparing for next Monday's 2.21.1
release is high, so right now is not the best time to be pushing
forward with discussing Sun Ray or multi-user terminal requirements.
These are understandably low on the priority list at the moment.  I
anticipate that multi-user servers in general will likely have issues
with the new GDM.  It is probably worth letting some of these general
issues get some attention before we start to worry about any needs
specific to Sun Ray.

> Why exactly?

Probably the main reason is that there are a fair number of users out
there who use Sun Ray on Solaris and Linux and who already depend on
GDM to meet their needs.

There are probably people out there with an interest in helping with
the work of getting the new GDM working with Sun Ray again.  We should
probably be encouraging such people with assistance and ideas rather
than shutting the door on discussion.

Furthermore, the existing GDM documentation claims that its interfaces
are stable.  So, if it turns out that we have been lying to people,
then I guess we shouldn't be surprised when some users get confused.

> So, I should just take your word for it and add interfaces where you
> say and how you say?  Even when you team has described large parts of
> your design as kludgy?  And you haven't decided how you'll handle
> ConsoleKit and HAL integration?

Perhaps you could offer some advice about how you think Sun Ray and
other multi-user servers (e.g. terminal servers) should integrate with
ConsoleKit and HAL in order to meet their needs?

Although ConsoleKit is now a hard dependency of GDM, it is not yet
clear to me if ConsoleKit is a good match for the needs of multi-user
environments.  My understanding is that ConsoleKit's main value is that
it makes managing multiple seats per-display more simple, so things
like VT switching is more manageable.  I know Sun Ray does not support
VT switching on its devices, and I would imagine that most multi-user
terminal environments also do not support VT switching.  Am I wrong
here?

So, I am curious what value ConsoleKit adds in such environments.
Perhaps you could explain?  Do you really need ConsoleKit to support the
sorts of things Sun Ray and other multi-user terminal servers want
to do?

If we want to use ConsoleKit to manage multi-user terminals, what
interfaces should people use to control the displays?  Is HAL the only
way that you imagine this will be controlled?  If I want to start a
display on a virtual frame buffer, for example, then how do you imagine
this working?  The Sun Ray case is, as I pointed out before, quite
similar to this question.

> I am not particularly interested in playing these types of guessing
> games.  We are moving forward the best we can to try to solve the
> problems that we can *see* and doing so by collaborating with our open
> source / free software partners.  Please join us and help us.

I do agree that Sun Ray requirements are probably not what we need to
be focusing on right now.  First, we need to get GDM working for
standalone desktop users, and there is plenty of work that needs to
be done to just support the more common use cases.

However, I don't think anybody is asking you to play any guessing games
or solve problems you can't see.  I do think that people at Sun are
wanting to explain our requirements and to get some feedback regarding
how we should approach getting Sun Ray working with the new codebase.

> It isn't only about the license.  It is about the process and the community.

I think GDM is one area where Sun, as an organization, has shown a fair
amount of community support and involvement.  Many other aspects of
GNOME have benefited from Sun's involvement.  The bulk of the
accessibility work in GNOME, for example, was spearheaded by Sun.  The
Sun Ray team has already told us that they are working towards
relicensing the Sun Ray software with an open license.

I am sure we can all do a better job of working together towards having
a healthy community.

> Have a beer - think it over.

Not a bad idea, it is Friday.

Here's to wishing everyone the best of fortune with the transition to
the new GDM rewrite which will hit the streets just next Monday!

Cheers and good luck,

Brian