GNOME Foundation / Mozilla Foundation Meeting Minutes Wednesday, April 21 2003

[Glynn Foster <Glynn Foster Sun COM>]

Hi,

We had a meeting with some representatives of the Mozilla Foundation
about how we could collaborate a little closer in future. See attached
minutes for details of the discussion.

Glynn

GNOME Foundation / Mozilla Foundation Meeting :: Wednesday, April 21 2003

Attendance
==========

Mozilla Foundation
	o Bart Decrem
	o Brendan Eich
	o Chris Hofmann

GNOME Foundation
	o Jonathan Blandford
	o Dave Camp
	o Glynn Foster [minutes]
	o Nat Friedman (:5)
	o Jody Goldberg
	o Bill Haneman
	o Havoc Pennginton
	o Marco Pesenti Gritti
	o Leslie Proctor
	o Tim Ney
	o Owen Taylor [chair]
	o Luis Villa
	o Jeff Waugh

Regrets
=======
Miguel de Icaza
Malcolm Tredinnick

Actions
=======
	o Owen to jumpstart a discussion of where GTK is going at a toolkit
	  level, and potential discussion points with Mozilla technology
	o Mozilla guys to hunt out a bunch of bugs and try and get the 
	  GNOME community more involved
Discussion
==========

	o Owen opened up the call, with the high level goal of how the 
	  the GNOME and Mozilla projects could work closer together and
	  without getting too involved in the technical details

	o Bart replied that they were excited to have the opportunity
	  of having this conference call, and that the more opportunities
	  we have of breaking out of our respective silos and working
	  together the better chance of success for open source and user
	  space.

	o Brendan spoke about the need for innovation, and not just 
	  clinging to web standards. We need to have the platform
	  more relevant, and believe there are pieces that are useful
	  for people readily available for a Linux offering. He spoke 
	  about the new for new application development to drive this.
	  He spoke about how we should move further with native widget
	  integration and user interface elements like print, open
	  and save dialogs. We should evolve this platform with the
	  help of GNOME and Linux, and others, to be more standard
	  compliant, low cost and open. It may not be as tight and
	  coherant as XAML/Avalon, but it should be competitive, and 
	  not something that all open source projects can build by 
	  themselves. Brendan stated the need for applications to be 
	  relevant, and not just a layout engine.

	o Brendan spoke about the needs of the Mozilla Foundation, 
	  primarily of testing existing web pages, of which many
	  bugs exist because of different standards, and gaps in
	  existing standards. We need applications so we can get
	  prompt feedback on these issues, and need wide distribution
	  for these applications.

	o Brendan was asked if he saw the Mozilla Foundation as providing
	  a platform. He replied that although there is a platform, it is
	  not widely used for it to become a competitive one. We need to
	  collaborate with other projects, and while we can't abandon it
	  completely, we would like to make it a part of a bigger platform.
	  Owen mentioned that the GNOME project was pretty much committed
	  to Gecko as a web rendering engine.

	o The web is currently pretty stagnant and is likely to change
	  greatly, especially if Longhorn penetrates the same as XP. In
	  2007 it's projected to have 30% penetration on Windows. That 
	  invariably means there will be XAML on the internet, and there
	  will be lots of pressure to get that ported to Linux. We don't 
	  want Mozilla to become a legacy content engine, so we need
	  to evolve web standards to be competitive and meeting the 
	  requirements of content authors. Mozilla would like to see 
	  alliances with others that we can ship, evolve and get market 
	  share.

	o Nat spoke about some developments at Microsoft - the 
	  merging of a bunch of teams for Avalon/XAML. Now have a
	  single team for web and native desktop rendering. GNOME
	  and Mozilla needs to align to counter this. One big 
	  fear moment was at the recent Microsoft PDC, with 
	  Amazon demoing their site written using XAML and pretty
	  compelling reasons for using it. Nat believes the options
	  are
		o Gain market share before Longhorn comes along
		o Clone XAML
		o Develop something new, cross-platform, and 
	 	  offering a rich web experience as well as a 
		  great native integration story, and getting
		  the big players involved. With this we have
		  a single library - the tarball is the standard. If
		  we're interested in a unified solution [GTK+ and XUL], 
		  how do we mix the two? How do we approach cross 
		  platform, without slowing down the schedule of each
		  project?

	  It was noted that Microsoft haven't ported many 
	  applications to XAML yet.

	  If we're going to be competitive, we need to follow the 
	  open source defacto standard route, that we're all working
	  on, rather than being bogged down with the standards process.

	o We need to slow the upgrade to Longhorn, and since that
	  is relatively costly to businesses, if we can make cross
	  platform applications work well, there is an opportunity
	  for Linux migration.

	o Bill mentioned the possibility of using SVG, which was never
	  completely implemented. One key thing for adoption is content
	  authors, and the need to make sure that content can easily 
	  created. There is a lot of opportunity to extend the spec.
	  Brendan replied that in his experience the SVG spec was
	  complex and not easy to change it. SVG has deviated from 
	  CSS/DOM standard, and if you wanted a mixed document you were
	  likely to run into trouble. He mentioned SVG doesn't have 
	  widgets, and re-implementing widgets using SVG primatives
	  would be unacceptably fat and slow for the foreseeable
	  future. Brendan asserted the need to leverage existing 
	  technology we have now.

	o Jeff asked how we should deal with duplication in technology
	  we have with GNOME and Mozilla today. Bart felt that it 
	  was a lot of work, for seemingly little gain. It was commented
	  the the model for Mozilla, GNOME and Win32 are substantially
	  different, with unification leading to lots of layers of 
	  abstraction. Jody pointed out there is a lot of potential
	  low hanging fruit - glib/nspr, pulling out unused/unimplented
	  APIs in nss and many incremental issues. Brendan pointed out
	  that there was little user benefit to redoing low-level plumbing
	  for small code footprint savings, and big opportunity for bugs
	  to creep in. Strings may have to be transcoded, and that may not
	  hurt much if the frequency is low.

	o Nat commented that we needed to focus this meeting, not on
	  technology, but more on user benefits. We don't need more
	  C and C++ code reworking. We need to build something that people 
	  actually use. Are we trying to 
		o Build new functionality into the toolkit
		o Allow people to build mixed web native applications

	  Owen asserted the need for a clearer idea of where we're
	  heading on the toolkit level and resolves to get a 
	  discussion going with GTK+. Owen has been trying to get 
	  some requirements from OpenOffice.org, Mozilla and SWT/Eclipse.
	  Owen will jumpstart a discussion, as a good way of getting 2 
	  silos communicating, since a unification plan is a little too
	  way off right now. There are many issues eg. Theme API that
	  can be discussed soon.

	o The Mozilla guys would like to see some sort of announcement
	  of an alliance with the GNOME Project. Jeff replied that we
	  would need to figure out direction first, since the board
	  doesn't specify the technical direction of the project. We 
	  would need to bring this discussion to the community, and until
	  then, this type of announcement doesn't really make sense.
	  Jeff mentioned that it would be good to chop these discussions
	  into singular pieces and bring to the community, also mentioning
	  it would be good to have some representation at GUADEC, OSCON
	  and OLS.

	o The discussion moved on to the Linux browser, with Owen
	  stating that the GNOME project defers these sorts of 
	  decisions to the maintainers, while vendors could override
	  any parts of the desktop - Sun and Red Hat have done this in 
	  the past.

	o Bart gave some background about the Firefox project, mentioning
	  that they had spent much time with IT companies building
	  relationships, and making sure they certify their application for
	  Firefox. Training costs are an important motivation for Firefox
	  development. Bart mentioned that GNOME has to do what is 
	  right for the project, but that it would be costly to have a 
	  browser war on Linux and asserted they were committed to providing
	  the best experience for GNOME users using the application. Luis
	  mentioned that any decision would be driven by the community and 
	  that all we could do as a board would be to encourage the benefits
	  of Firefox.

	o Jeff spoke about the options the Mozilla Foundation might have
	  if they really wanted to pursue the browser issues immediately
		o Convince all distributions to ship Firefox
			- if done privately, community may feel that
			  Mozilla has ignored well tested community 
			  processes
		o Convince Marco himself. If Marco wants to work
		  on Epiphany, then there is no stopping him, and he
		  has an awesome group of people to help out.	
		o Improve the level of visibility and interest in Firefox.
		  The application is used in the community by a few 
		  individuals, but not really on the radar of the GNOME 
		  project.
	  
	  Jeff mentioned that Firefox and GNOME aren't perfectly 
	  integrated - in terms of user interface guidelines, preferences, ..
	  Jeff suggested if it was possible for Epiphany to become the
	  official Linux port of Firefox.

	o Bill asked for confirmation that work done by Sun, IBM and
	  others will be integrated as soon as possible into Firefox
	  and whether it was still valuable. Brendan confirmed.

	o Marco felt that it was important for joint collaboration, and
	  while both projects had somewhat conflicting goals, there was
	  a lot of space for commonality. Marco mentioned that we should
	  be focused on a better solution for uses when compared to 
	  Microsoft technologies, and that it was a shame that these
	  types of discussions had taken so long to start. Marco felt
	  that Epiphany was a better solution for GNOME, not because
	  he wrote it, but because it met the needs of users and was more
	  inline with GNOME's direction. He felt that the toolkit is 
	  only one of the problems with an integrated design for the 
	  whole desktop, and that many important things are sacrificed
	  for the sake of cross-platform.