What is freedesktop.org

[Havoc Pennington]

Hi,

I think a FAQ is a great idea.

The major point perhaps is that it's an open source project and thus
whatever the contributors make of it. Which has two aspects, one is
anyone can have a say, and two is you can only have a say by
contributing - easiest contribution is to get on the mailing list and
participate in threads.

But that said I'd like to outline the _intent_ as _I_ see it, and maybe
I can clear up some of the issues from the IRC log posted. If what I
have to say doesn't match what it says on the site, it's a Wiki after
all, let's fix it up. Some KDE devs have root access on the machine that
hosts the site, and many more have Wiki account access to change the
site.

I'm by no means the largest fd.org contributor at the moment though, so
I don't feel what I say is even the largest vote anymore.

So -

My original intent in 2001 was to host draft specs and conversations
about specs to be shared between desktops.

Over time this started to include hosting some software, e.g. pkgconfig
and the CSL sound stuff. I posted the following this Summer proposing a
some changes in direction:
https://www.redhat.com/archives/xdg-list/2003-July/msg00101.html

The IRC log Lubos posted had a lot of discussion of what is standard and
what's just hosted. A couple points on the idea here:

 1. fd.org doesn't have an IETF/FSG/etc. style standards 
    process. This is what "standard" really means to me. 
    
    In absence of this process the only meaning of "standard" 
    I care about is "GNOME and KDE both use it," honestly.

    Previously there was discussion of standard vs. not and 
    in response we grouped the specs on this page into categories:
    http://freedesktop.org/Standards/Home
    as it says there, if something is in the wrong category just 
    bring it up (or use the Wiki to change it yourself - if you 
    don't have permissions and are a developer, just ask on 
    #freedesktop or mail the list to be added to permissions - 
    the permissions system is just to avoid defacement)
  
 2. The reason we host anything "on topic" including competing 
    projects is that we want to allow natural collaboration and 
    let conversations start, without having to decide between 
    unfininished projects in advance. Also competition is good
    so projects actually go get buy-in from GNOME and KDE.
    
    So right now there are 4 or 5 X server projects on there, 
    and it looks like we'll have multiple input method projects
    and sound/media projects.

    We simply don't have a mechanism to choose among unfinished 
    projects that neither GNOME or KDE are using yet, but it 
    seems like it's valuable to give them a place to evolve.

 3. The idea of the "freedesktop.org platform release" is to 
    define the set of stuff that GNOME and KDE are both 
    _using_ - stuff that's already adopted. And make it available 
    as a coherent/tested set for people to download. So instead 
    of telling people to go get pkgconfig 0.15, Cairo X.Y, 
    fontconfig X.Y, etc. we just say "get freedesktop.org 1.0" just 
    as we'd say "get KDE 3.2"

    However, this set has to be conservative and confined to things
    we've actually agreed on.

    So this release would be what really defines what's "standard" 
    (where "standard" means it's widely used and useful, not that 
    any standards body was involved)

    There are some things both of us _already_ depend on that 
    make a lot of sense for this release, IMHO.

    Some have suggested e.g. Xserver in the fd.org platform release; 
    I don't think that makes any sense until we know which X version 
    the major distributions will include.

 4. What to do if you don't like stuff on fd.org - I think what you 
    do in this case is the same as if you don't like some code in KDE. 
    Get on the lists and point out the technical problems, why it won't 
    work for KDE, and propose what would work. Engage with the 
    discussion and work it out.

    Sometimes in open source you have to fork, and so create a different
    project on fd.org with a different proposal if required.

    Remember fd.org is just a bunch of individual contributors, they 
    all have different biases; some of them are more technically skilled
    than others. I'm sure you're all used to having both great 
    contributors and questionable contributors on a project and working 
    around the dubious ones. This is just the open source process.

 5. I recognize there's a big macro question of whether KDE should 
    support an effort like fd.org at all; I've tried to make the 
    case for doing so at Nove Hrady and elsewhere.

    I hope though if you don't want to support the overall goal, just go
    ahead and say so; there's no point nitpicking the details. If you do
    like the overall goal, I hope you _will_ nitpick the details until 
    they are right.

Anyway, just a lot of poorly-organized thoughts, but I hope it makes
some of the ideas clear. Ultimately nothing is set in stone, it's up to
GNOME and KDE contributors to shape what gets done in our shared
collaboration space.

Havoc