Bringing fdo.org to the next level

[John (J5) Palmieri]

On Fri, 2005-04-15 at 07:13, Ely Levy wrote:
> On Fri, 15 Apr 2005, Philippe Fremy wrote:
> 
> >
> > Well, it seems that so far, freedesktop has worked very well in an
> > informal way, without formal board and approval process.
> 
> freedesktop didn't do anything, it just hosted some project and one
> successful mailing list. Standard discussions were before fd.o was around.

And this is what made it successful.  It became a single point of
communication where people of different camps felt safe to discuss
contentious issues.  It became a demilitarized zone of sorts and let the
projects evolve on its own merit.  By any measure it is hugely
successful because it did the one thing a formal standards organization
does not do well.  It created dialog. 

> > The discussion here are usually enough to foster technical discussion
> > and reach consensus between desktops when consensus can be reached.
> 
> But why be so passive?

Because it works.  Trying to force standards in this case will create a
platform of just good enough software.

> Why not activly trying to reach it?Why not serve as a mediator?
> and what about the users?how would they know which desktop supports which
> standards in case they decide to use only those who works well with each
> other? and about programs who wants to tell you if you support fd.o 1.3
> then our program would with with yours instead of listing 10 different
> semi standards.

Why does a user care as long as the desktop just works?  This is a
developer issue but at this point not even D-Bus is mature enough to
create a platform that is not tailored to a particular distribution. 
BTW there will be a fd.o 1.0 platform release at some point from what I
understand. 

> > Can you explain what you think the current problems are with the current
> > organisation, and what a new organisation would do to resolve those
> > problems ? Your proposition sounds a lot like LSB to me and it was clear
> > here that this is not the way freedesktop wants to work.
> 
> There is no origanisation. I have no problem with how fd.o is now,
> a hosting site. I'm trying to show the advantages that org which work
> more activly like OASIS or LSB can have. We need one which would be
> dedicated to free desktop.

If components in freedesktop warrant standardization they will get to
the correct body and be stuck in comity for years.  Have you read my
rant (http://martianrock.com/?p=43)?  A snippet:

Fd.o was conceived specifically NOT to be a slow working standards group
formed of committees that put out specs that were relevant years ago but
somehow seem outdated and over engineered when they are released. Fd.o
IS A proving grounds for specs and implementations that are of interest
to unifying the many different Linux desktop platforms with some common
infrastructure. Projects still have to make it on their own merit. If a
specification becomes a standard it is because all or most of the
interested parties have agreed to use it and not because it is hosted on
fd.o.

> > If your proposition relates to the lengthy discussion about d-vfs or
> > dconf, my very personal opinion is that you should first start to write
> > code, and then show it and discuss it (like Linus said, "show me the
> > code!"). Freedesktop was created in the spirit of people writing code
> > and discussing together real problems. It has worked in this way so far.
> > I consider the two recent discussion a major exception to this and I
> > hope that this is not going to become the trend.
> >
> I don't care much for dvfs sorry:)
> Writing code is a neccery step but it doesn't have to be the first one.

Short, concise specifications are fine to start with and give a project
some direction.  Start to complicate things by putting the cart in front
of the horse and nothing ever gets done or you end up with a system that
strokes the designers ego but does nothing for the user.  

> Actually that exactly what this mailing list is ment for to talk before
> writing, to do a bit planning to maybe be able not to write 10 different
> protocols and use one instead. To see how non realistic it is
> just think what would have happen if internet had 10 version of tcp...

But there were many different protocols that competed with each other. 
TCP/IP just won, and not on discussion alone.  TCP/IP was implemented
before the Internet became a reality.  

> > Write code, show it and KDE, Gnome, XFCE, mozilla and other people will
> > tell you "This is crap, I am never going to use that" or "this is great
> > stuff, I'll convert my code to it next week" or "good work but you need
> > to add this and this before I look at it again". In my humble opinion,
> > there is no point in discussing the topic anymore. It will be quicker to
> > rewrite a kconfig for a new configuration backend, than to discuss
> > whether this is possible, desirable and what the possible problems could be.
> 
> Yea this is one way, that it's not the most efficient one by far.

Actually, it has been.  No other organization has been able to get these
historically hostel groups working together.

> This way takes years dicussing with one project going to the other
> making changes going back to the rest of the projects making changes
> again.

And forcing standards works faster?  I think what you describe above is
the evolutionary process of software development in general.  I don't
think you can go any faster without compromising the robustness of the
standard.  By standards term fd.o has been at light speed if measured by
rate of adoption.  

> IMOH freedesktop contributed to the desktop mostly cause no one knew
> exactly what it is. People did look at the standards here as standards
> and went to code them. If gnome/kde/XFCE people think a standard would do
> good for their desktop they would write the code for it. That's the point
> I'm trying to make if you had a org which activly encourage people to
> write standards and use them it might contribute a lot more than just
> hosting a mailing list.

fd.o provides a natural Darwinist approach to standards which allows for
fast innovation while filtering out bum standards.

> > In the free software world, the consensus is reached upon code, not upon
> > discussion.
> 
> That what got us into not being able to have desktops which work with each
> other in the first place. That what this mailing list tried to change.

No, it got us competing desktops that pushed each other to become
excellent in short periods of time.  There were trade-offs which fd.o is
trying to fix but nothing is perfect.  We already have true standards
organizations, we don't need another one.  I think fd.o is working fine
the way it is. 

--
J5