Re: A question for all the candidates.

[Daniel Veillard <Daniel Veillard w3 org>]

[ Dohh ,,, it really grew longer I expected, I hope there won't be too
  many questionnaire like this one, Daniel ]

On Mon, Oct 16, 2000 at 06:22:38AM +1000, Martin Sevior wrote:
> 
> Dear Candidates,
> 	        One of the duties of the Gnome Foundation is to provide
> "Direction" for the project. In a Free Software environment like Gnome it
> is not immediately obvious how this will be achieved. I can imagine
> several ways in which direction can be provided in terms of strength and
> nastiness. Here are some of them:
> 
> 1." This project is wasting resources. It will be removed from Gnome CVS."

  Bahhh no ... whoever put time and effort in code would be pissed off,

> 2. "This project is wasting resources, we recommend that no distribution
> (especially Helix-Code) include it in their release". 

  Practically you can't do it. What is clear is that the Board is the
instance saying Gnome-X.Y includes packages A version a , B version b,
C version c ... A distribution is very welcome to make additions to it
it is substractions which would be a problem. Also releasing "advanced
betas" is a good way to recruits more volunteers and get more bug reports
which are precisely the things needed to get out of beta :-)

> 3. "Project X and project Y have the same goals. Have you guys considered
> a merger"? 

  As a board member I would definitely do that ! "considered" is the
right term I would not force them but certainly suggest.

  Let's take a practical example for what seems to be your concerns
in question 1/ 2/ and 3/, HTML widget. There is gtkhtml, the Mozilla
code and gtkhtml2 develooped by Anders Carlson (and others):
  - I definitely tried to get andersca and ettore to agree on goals
    and get them to work on the same code base avoiding duplication
    of effort. I thing as a Board member I would do the same.
  - Considering it was clear both project could not agree (whatever
    the reasons) and assuming gtkhtml2 is free software, then I have
    for months tried to get gtkhtml2 code into Gnome CVS. I was told
    it would be done soon.

  For me it makes sense to let time decide what code base will emerge
on the long term, and since Bonobo can be easilly configured to use
one or another rendering widget depending on user preferences or the
page viewed (Nautilus allows this IIRC) that selection will be done
naturally. If the code ain't used it won't get bug fixed and will end
up being obsolete. If all 3 widgets have a decent user base, then they
they can all live simultaneously !

> 4. "We really think project Y should be supported. Gnomers please hack on 
> it." 

  Sounds like an alarm bell ... I would expect that message to be
carried by technical discussion lists or user feedback. However
I see the Board asking for some infrastructure works like I said
when answering Alan message earlier, for me asking the membership
to find a solution to the mailing lists and archive problem is relevant
(yes I consider this a problem, at W3C we maintain list facilities and
I know how useful they are).

> 5. "We will use our money to provide a prize for the best patch of the
> week to project Z".

  No as Maciej pointed out this would be an unilateral technical
decision. A prize for a new logo for a Gnome conference on the 
other hand would be fine IMHO because it benefits the whole project.

> 6. "We think this project is so important we strongly suggest that Gnome
> companies consider hiring someone to work full time on it."

  Hum ... First I'm afraid this would not work well. Second if
something is clearly lacking again the alarm bell will have rung
in the technical forums a long time before. If some companies
think it's important they should not (and won't anyway) wait for 
the Board to take care of it.

> 7. "We think this project is so important the Gnome Foundation will hire
> someone to work full time on it."

  No way !!!
  First we would depend on the money stream for technical devel and
I strongly oppose this by a matter of principles. Second this would
again be an unilateral technical decision. Third the relationship of
that developper "selected by the Board" in the project may actually
not help that much. GNU used to have (I don't know if it's still true)
a person employed to work on gcc, I think this led to some problem
when patches were not considered applied fast enough. Working for the
project in that case may end up being very frustrating for both that
person and for the other people depending on the outcome of that work,
sounds a really bad idea ...

> So candidates, these are some of my ideas about how direction in Gnome can
> be achieved. (I'm not saying I agree with any  of these methods.) 
> How do you think direction of Gnome should be achieved? 

  I have found the description "Meritocracy of Geeks" applied but
to another project. I think it reflects the current state more or less.
Geeks buy into technical arguments and are usually impossible to led
into a direction they don't want to take :-) . Any attempt at forcing
a direction will simply be ignored if not backed up by strong technical
arguments. And those arguments won't be discussed at the Board level 
but in the lists (one more reason to have good lists support !) or IRC.
So I don't see the board deciding directions by itself, but could help
by :
    - spreading the information when it's clear that a technical
      choice has been made, i.e. it doesn't do the decision but try
      to get it promoted, recorded, archived, etc ...
    - when there is an open choice and that is serious enough,
      compile the pros/cons of both side, and try to keep the
      information visible so that people understand the issues

  I don't think of the Board as anything like a legal institution,
for me it's role is at most advisory on technical stuff. There is 
a few things for which it has a deciding role, like conferences, and
releases, but I expect those to be based on consensus decisions and
when there is an opposition to a decision it should be recorded. 
If we end up not having consensus on too many cases that would prove
serious troubles for the Gnome project itself (unless the Board
membership just can't work together, then it will be simply ignored :-)

Daniel

-- 
Daniel Veillard w3 org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux XML libxml WWW
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | Gnome rpm2html rpmfind
 http://www.w3.org/People/all#veillard%40w3.org  | RPM badminton Kaffe