Conference call summary, 23 April 2002

Participants:

Chris Blizzard			  Keith Packard	   
Alan Coopersmith		  Havoc Pennington    
David Dawes			  Branden Robinson    
Egbert Eich			  Leon Shiman	   
Marc Evans			  Owen Taylor	   
Marc La France			  Mark Vojkovich	   
Jim Gettys			  David Wexelblat    	    
Mike A. Harris			  Carl Worth	   
Rich Murphey	   		                     

The conference call was long; and at times a bit chaotic, so what I've
tried to do below is identify the important themes of the conversation
and the major views expressed on those themes rather than a linear
summary of the discussion as it happened. Attributions are approximate;
unattributed points do not necessarily reflect general consensus.

What's not working
 - Lack of good organization for communication? (Leon S)
 - The problem is trust (David W)
   - Personality conflicts?
   - Factions with different goals?
 - The technical goals of the different "factions" are similar (Havoc P)

How does the XFree86 project view itself in relation to the
 larger concept of the X window system?
 - XFree86 is exclusively the layer beneath the desktops (Marc LF) but 
   does need to interact with the desktops (Egbert E)
 - People tend to perceive XFree86 as the defacto standard
   implementation of X (Branden R)

How does new development happen?
 - Traditionally this has been a role of X.org (Mark V) but pay-for-say
   doesn't work in OSS (Jim G). Maybe X.org processes can be adapted to
   address these concerns?
 - Standardization shouldn't be a separate process from development;
   code collaboration is needed.
 - Should be a single path for adoption of technology so there isn't
   gratuitous duplication. Not clear currently who one approaches if 
   one wants has an interest in enhancing X. (Havoc P)
 - Render and RandR were examples of new technology being adopted with
   the current process, but things didn't always go smoothly. (Much
   discussion.)

Separation of drivers from releases
 - There is considerable interest from distribution maintainers (Branden R,
   Mike H) in seeing more rapid release cycles for drivers.
 - Most of the work to enable this was done with the modularization of
   drivers in XFree86 4.0; examples of Nvidia ATI releasing drivers
   independent of XFree86.
 - Some discussion whether further changes, for instance breaking out
   the drivers from the XFree86 release would help speed up driver
   development.

What is the role of xwin.org?
 - A place for communication between users and developers (Mike H)
 - A place where development that is free to fail can go on (Branden R)
 - xwin.org is a broken idea; X development needs to go on with active
   cooperation in a single forum (Havoc P)
 
Could experimental development go on in the XFree86  repository?
 - Is there reason for doing that rather than in, say, sourceforge?
 - Extensive discussion of whether DRI would work better in
   the XFree86 repostory or not.

Moving forward from here 
 - Discussions of vision, structure are basically a distraction;
   nothing is blocking people from just doing work.
 - [email protected] is a better venue for communication 
   than conference calls