Re: Draft of Proposal for the GNOME Foundation.

[Nat Friedman <nat helixcode com>]

Havoc Pennington <hp@redhat.com> writes:

> IMHO we will have unfortunate signal-to-noise ratio if we do that. The
> list is open to anyone who wants to subscribe, of course, but it's
> only been announced on gnome-hackers and that's probably a good idea
> (I think), combined with inviting people personally if you know they
> should be invited.

Ok, I wasn't talking about inviting everyone into the list.  I just
wanted to solicit comments from the people who make up the project.
Yeah, gnome-devel-list would have been a better choice.  Maybe I
should send it there?

> This should have at least been discussed before you did it, because
> there was agreement previously not to post there.

I didn't know that.  I must have missed that mail.  I'm sorry, had I
known, I definitely would have waited for people to reply.

>  - decisions don't get made in 400-person mailing list discussions;
>    they were originally made by Miguel and some advisors, and more
>    recently the steering committee ends up doing it. You can of course
>    have a discussion on the list, but in practice a huge group can't
>    make a decision. (You say later that decisions just "happen", I 
>    don't think that's really it; I think rather a small group 
>    makes the decisions, and at some point everyone else decides
>    to go along with it.)

You're right, that was somewhat inaccurate.  But I do think that, once
you start creating /formal groups/ whose assigned task is to make
decisions, you have to be a lot more careful about keeping it open,
public, and observable.

A lot of decisions have been made on IRC, which, while admittedly not
a sea of 400 developers, does not present significant barriers to
participation.  A formal group assigned to make release engineering
decisions might opt to use their own private channel to discuss the
issues at hand, instead of letting things occur more naturally out in
the open.

You're absolutely right, though, I was a bit too emphatic here.  I'll
tone it down.

>  - we have some things that have to be private, such as companies 
>    interested in GNOME. Right now, these companies talk to other
>    companies, like Red Hat or Helix, and non-corporate developers
>    have no opportunity to participate in private discussions.
>    This is what will keep happening unless we have a board that 
>    can have private discussions; at least with the board, everyone
>    in GNOME has a chance to be on it. 

Yeah, that's the role of the board.  A company can approach the board
and engage in private conversations.  I totally agree.

> I'm sure you agree here, but you might want to tone down the
> principles a bit, or we'll get people saying that the board
> contradicts our principles.

Yep.

> >        Free software  licensing has always been  a mainstay  of GNOME,
> >        and we must     ensure that this   tradition  continues.    The
> >        foundation must not allow any software module  to become a core
> >        GNOME  component unless  it  is licensed   under the GPL,  or a
> >        GPL-compatible license.  GNOME should strive  to be free, while
> >        still being friendly to ISVs and commercial developers.
> >
> 
> This paragraph is meaningless unless you define "core component",
> which Miguel has refused to do and your above stuff about elitism
> would make impossible. ;-)

Yeah, strange, I thought I'd postedited that, actually.  I meant to
make a stronger statement:

    The foundation must not allow any software module to become a
    GNOME module unless it is licensed under the GPL, or a
    GPL-compatible license.

The straw man has been hung -- start hitting.

> It sounds like our only definition of "what is contained in GNOME"
> will be the contents of the releases; I propose that we require them
> to meet the open source definition (or the FSF equivalent), and leave
> additional criteria up to the judgment of the board. For example if a
> license is technically open source but certainly counter to the spirit
> of open source, as many corporate Yet Another Public Licenses are, the
> board could reject that license.

The idea -- working off a document/specification which defines the
notion of "free enough" for GNOME -- is good.  But I think that the
FSF has already done this for us.

> To pick a concrete exception to your rule, I think we should be
> willing to make the Mozilla module of Nautilus part of GNOME, and it's
> not GPL-compatible, but the MPL is overall a fine license designed in
> good faith.

I guess I don't agree.  But I don't want to dive into the
licensing-debate rathole... I'll leave that for someone else to
tackle.

> I do suspect that the GNOME foot and the word GNOME are already far
> too widely-used to trademark meaningfully, but I guess many people
> acknowledge Linux as a trademark though it's in the same situation.

Yeah, we could sic some IP lawyers on this, I guess.

> Isn't the chairman typically just the person who enforces Robert's
> Rules of Order? That sounds like something Miguel would not want to
> do. ;-) He's welcome to it of course.
> 
> I don't think we'll have a "chief GNOME dude" position really.

You may be right that "chairman" is entirely the wrong position for
Miguel.  But I think that it makes sense to give Miguel a special
position such that his leadership will be recognized and leveraged on
an on-going basis.

> >     approval  of the board  of directors.  Debian and  the FSF will be
> >     given permanent positions in this body.
> >
> 
> One worry about this is that this forum will certainly require
> nondisclosure agreements or at least promises of nondisclosure, and
> the FSF has a history of not wanting to sign those. 

Good point.  Perhaps we will want a forum and a separate advisory
board.

> BTW I think board members will have to be willing to agree to keep
> information confidential when that is appropriate. We should make it
> clear that that is a prerequisite for serving on the board.

Another good point.

> >     A   referendum   can be   issued by    any member  of  the general
> >     membership.  
> 
> We need to require some sane number of seconds to the motion to
> vote. Probably a percentage of total membership, say 5%. Otherwise
> we'll have a vote every other day.

Well, I figured that we'd let the software handle this.  The thing can
hold multiple running referenda, and if no an insufficient number of
people vote on the thing, that means it's pretty likely the requisite
5% wouldn't have seconded it.  Unless voting and seconding are
fundamentally different actions with different behavioral
characteristics.

> 7 days is most likely too short, people often take a vacation for a
> week. I'd rather see 2 weeks.

My concern is that we want to avoid an organization which throws sands
into the cogs of progress... but probably it takes at least 2 weeks
for the project to come to any major decisions now, anyways, so this
seems pretty sane.

> >     The size of the board will scale with the number of modules in the
> >     project.  The ratio (or whether or not this makes sense at all) is
> >     an open question.
> >
> 
> As Alan says I think the board should be a fixed size. Beyond 9 or 11
> people, it will be too big to get anything done.

Yes, it's true.  The question is: what do you do when the project is
so massive that 7 people cannot necessarily represent every aspect of
it?  You'll effectively need professional board members, it seems to
me.  Or a deeper hierarchy to ease the communications problems.  Or
some really unique individuals.  Maybe one of those is possible.

> >   The  board of directors   will  be responsible for   authorizing the
> >   release of a new version of GNOME.  The board will determine the set
> >   of  modules which will  make  up the release   at  least 60 days  in
> >   advance of the  release date, subject  to unanimous  approval of the
> >   module maintainers.
> >
> There should be a provision to punt modules that are initially planned
> on but just don't work out for some catastrophic reason. (For example,
> the company writing the module goes bankrupt, or decides they don't
> want to do it, or the module turns out to be buggy beyond all
> redemption.)

Yeah, I figured this would just sort of happen, with or without
policy.  But we can mention it explicitly.

> >   If a new module is being included in a release, all its contributors
> >   have the option to become part of the General Membership.
> >
> 
> It isn't clear to me whether you're proposing that a module must be
> planned for a release before its contributors can become members. 
> There are plenty of people on gnome-hackers writing stuff 

Ok, I was thinking about a situation like this:

    Company XYZ has a spreadsheet product called Add.  It's been out
    for 2 years.  Suddenly they decide to open source it.  They
    approach the GNOME foundation board and make a presentation.  The
    board says "Great, we'll include that in the next release.  Now go
    open source it, etc etc."  Add becomes Gnome-Add.  Now, all of the
    people who contributed to Gnome-Add, who previously had nothing at
    all to do with GNOME, are free to become members of the
    foundation.

For people on gnome-hackers, committing stuff into CVS, I wasn't
proposing that they be excluded just because their modules aren't yet
release-worthy.

> >   The board of directors will be primed by the election  of a slate of
> >   initial members.  Anyone   may propose a   slate, so long  as it  is
> >   approved by at least 5 general members.
> >
> 
> Let's say a percentage, not a fixed number. Maybe 5%. 

I was working off Bart's number from before.  But a percentage does
make more sense in a body with unlimited size :-).

> Also we need to put in one of these voting systems as previously
> discussed (so you have to get a majority to win, etc.)

There ought to be some software packages out there that do this; I
think I saw some a while back.  One remaining question about votes of
the General Membership is whether or not their votes are private.  The
Board votes clearly have to be public, and it's probably the same with
the GM, but I'm not sure.
 
> >     1. Historically, decisions in GNOME  haven't  been  made.  They've
> >        just sort of happened.   Does  this foundation really  have any
> >        chance of doing  anything useful except issuing press releases?
> >        If it's going to be  useful, I think it's  pretty clear that it
> >        has to be a natural growth step from whatever "organization" we
> >        have now, and it has to be wide, wide open.
> >
> 
> As I said above, I think the board etc. is actually more open than the
> way we've decided stuff in the past. 

I hope you're right!  That'd be awesome.

Nat