Re: polarization

[Havoc Pennington <hp redhat com>]

bruce perens com (Bruce Perens) writes:
> OK, thanks, this is better than the silent treatment I seemed to be getting.
>

I think people are just reluctant to "go negative."
  
> This is not necessarily a reason, alone, for election to the board, though.
> You can do speeches even if you don't represent something. It's just harder
> to get the show to pay for transportation and lodging that way. So some
> good speaking opportunities for the project go by the wayside.
> 

Usually we solve this by trying to get someone especially appropriate
for the event, e.g. Owen to talk about i18n, or Martin for shows in
Germany, etc. and distribute the load that way. Miguel does the big
shows such as LinuxWorld. People seem OK with a substitute for Miguel
if we can give a good rationale why the substitute is good for their
show.

Honestly though, talking at shows isn't one of the problems that I
think we need a foundation for; we're appearing at the important
shows. And people want an active contributor to the project to give
the talk, so they can ask questions (compare Bonobo to KParts, why
does gmc suck, etc.).
 
> I want to hire documentors to write books. We don't have enough
> books that are themselves Free Software. You of all people should
> know about the need for that, and unfortunately it's not always
> possible to get a publisher to pay an advance for writing a book
> that is itself Free Software. There have been some big changes at
> New Riders and other publishers since you did your book, and it's
> not so easy any longer. Look where Laurie is working these days.
>

It's a possibility, yes. However, I would rather see member companies
finance members of a documentation team. One problem with a book like
mine is that it isn't actively maintained; what we need more is an
active documentation team that's constantly keeping things up-to-date,
and has lots of people involved. The GDP is pretty good at this,
though they haven't attempted a book yet.

One vendor that recently asked to join the foundation offered to hire
a person to work on GNOME, and docs was a suggestion we had for what
that person could do. Red Hat funded the GNOME 1.0 User Guide and
Helix and Sun both have people working on docs, I believe.
 
The foundation itself would not be qualified to contract an author
directly to write a book; there are lots of administrative issues,
you'd need an editor, and you'd want better distribution channels than
the FSF has. So I would want to see us work with a publisher of some
kind if we did this.

Anyhow, a complicated issue. 

> Havoc, the foundation is about money. It's not about governance of
> the project, that belongs in the developers hands. Liason with
> corporations is about money.  Some of the things we do other than
> coding take money. That's the number one task of the board, whether
> or not they know it today.

I don't see how we can have a foundation about money that's separate
from the developers' governance of the project.  This is by no means
my impression of our motivation for creating the foundation. GNOME has
been doing just fine for 3 years with no money of its own to speak of
(though millions of dollars worth of code, hosting, etc. have been
contributed). In fact we had and still have substantial disagreement
on this list about whether the foundation should have any significant
funds at all. The archives will have that discussion, you probably saw
it.

The interim steering committee has been playing the role of the board
for half a year now, and I think that's a pretty good indication of
the kinds of things the board will likely get involved in.

To me the foundation is a mechanism by which the developers are going
to govern the project. For example, the board will be in charge of
coordinating releases. Basically GNOME is a huge project; and while a
project such as Debian is of similar size, it's much more
parallelizable than the core components of GNOME are. We have much
greater intermodule dependencies, such that 100 people might have to
be talking to each other daily to get a release out. Witness how much
the Debian guys like to joke about the dependency lists of the GNOME
packages.

What this means in practice is that we need a smaller group to
maintain communications and keep things working smoothly. The board is
not making technical decisions, but it is making sure such decisions
get made, and that they get communicated well between relevant
developers.

GNOME will have to deal with money, yes. We'll have to deal with
corporations and balance their interests. But these aren't really the
primary tasks; these are means to an end. The primary task of the
board should simply be to help GNOME run more smoothly, solving
problems as they appear and as required. We want to coordinate the
technical efforts of the involved individuals and corporations; we
don't want to replace or control those efforts.

That's why I say close familiarity with the project is important,
because I see the board as a body that works with developers.

Whenever a decision is public, about money or not, the whole project
will know about it and comment, with the board trying to extract a
consensus. When something absolutely must be confidential, the elected
representatives of the project will have to make the best call they
can. But only when absolutely necessary. I don't think it's correct to
say that the board handles money and the developers do not; that's not
the way we've set it up.

Corporate involvement up to now has mostly been them funding one or
more developers, and the developers then interfacing to the project as
individuals. Our model here is very different from an industry
consortium. We should keep it that way.



We aren't talking about naivete and unawareness of the ways of the
corporate world here. Many of the core contributors are employed by
free-software-related companies. We all know how GNOME fits in to the
making money side of things for our companies; we do talk to the
business guys. And at Red Hat for example, we also have pretty
substantial insight into the kinds of things our customers (who are
mostly companies as well, some of them large) might want. Someone like
Jim Gettys who's edited the HTTP spec for the IETF and been involved
in the X Consortium has certainly seen companies involved with
nonprofits before; ditto for Daniel at W3 and Bart (see
http://www.decrem.com/bio.html).

I do think the situation is fundamentally different from the one two
or three years ago, when the term "open source" was invented and we
had a lot of trouble talking to business. Now half the community works
for free-software-related businesses. We have nontechnical people with
substantial business experience with a stake in our success.

Moreover, we're no longer begging; we have a strong bargaining
position and companies are coming to us on our own terms. We don't
need to woo them, we just need to help them work with us.



I would say the board should just be the logical extension of the
GNOME Project. The formation of a nonprofit corporation is a legal
technicality we need for various reasons. It shouldn't take on a life
of its own and we should fight to avoid it doing so.

If the board becomes primarily about money I think we're in big
trouble. Commercialization of GNOME should be left to companies; money
should also be left to them when possible. This keeps the nonprofit a
servant of the project and the developers, rather than a competing
entity with a life of its own. A truly independent nonprofit
contributing the resources we get from member companies today would
require a few hundred million endowment; we aren't going to get
that. And the whole point of free software is that we don't need to.

> I fear that in the future the foundation will also be about legal defense.
> Of course, that's money too. I'm hoping that our new friends will remain by
> our sides through patent lawsuits and so on. That's probably the true test of
> them.
> 

All the more reason to avoid depending on a foundation that has large
coffers - if we don't have an interesting amount of money, we can't
get too upset if they sue us for it. The GNOME Project should be
sufficiently independent of the foundation that it can continue
without too much pain regardless of pending lawsuits against the
foundation.

Of course, a sufficiently hostile entity could go after member
companies or individuals. But I think that's an exceptional crisis,
certainly not the day-to-day business of the board, and I have no
doubt that we'd be able to help individuals that needed it.

Havoc