Re: gnome-vfs dependency for bonobo (was gnome_mime vs. gnome_vfs_mime)

[George <jirka 5z com>]

On Sat, Dec 02, 2000 at 06:20:48PM -0500, Michael Meeks wrote:
> Hi George,
> 
>         I concur broadly with much of what you say; however the reasoning
> at the end worries me:
> 
> On Sat, 2 Dec 2000, George wrote:
> > 2) GNOME 2.0 should be out in 2001.  If we want to keep the support of
> > companies like Sun or even keep existing users, we must provide new   
> > stable releases.
> 
>         I think a constant stream of new 'stable' releases is not good for
> anyone, I am suprised that you want to get Gnome 2.0 out quickly, surely  
> it is more important to be correct than fast. A lot of work will need to  
> be done to Bonobo to start merging it with UNO, and I in no way feel that
> this should be rushed in any way whatsoever.

I want GNOME 2.0 be correct, but not neccesairly perfect.  We have slipped
heavily into a Debian/GIMP like release policy.

I don't think things need to neccesairly be rushed.  But not everything has
to happen in 2.0.  There WILL be releases after that.  We should pick the
things that really need to be done and do those.

I used to be of a different opinion before, but after our performance with
1.2 and 1.4 slippage, I don't think this is a good idea.

>         I think it is of vital importance to get the core infastructure as
> good as humanly possible, and then release it and commit to not changing  
> it at all and providing as much backwards compatibility as possible for
> ever. This militates against very rapid release cycles.

This will not work.  Mostly because by the time you get to be "perfect"
you'll realize that some of your stuff is completely braindead.  Something
that you thought good and perfect just a year ago.  Yes we should strive to
have a good base, but we are missing too much.  You cannot build a large
building like GNOME if you spend all the time on the foundations.  In reality
a user doesn't really care if something inside is an utter hack and ugly as
long as from the outside it all works.

I just don't want GNOME to end up like Hurd, which by the time it will be
usable by end users will be old and obscolete technology.  Do you think Linux
si doing things perfectly?  Naaah.  It's actually quite crappy.  Yet it's
getting better incrementally.

I would do so many things differently if designing gnome from scratch again.
But we do not have this luxury.  I hope that in like 10 years I'll start work
on a competitor to GNOME which will do things the "RightWay(tm)".

I'm also looking at this from a perspective of free software.  We MUST build
something that is usable by end users asap.  We must get a critical mass
where new software will automatically get written for the GNOME platform.
We must get free software onto people's desktop.  I don't see the "we must do
things perfectly even if it takes a long time" in that equation.

We must also notice that we are in competition with the rest of any desktop
software, be it kde, windows, macos, whatever...  people will not use us if
the stable release will always be obscolete because our release cycles are
too long.

George

-- 
George <jirka 5z com>
   I thoroughly disapprove of duels.  If a man should challenge
   me, I would take him kindly and forgivingly by the hand and
   lead him to a quiet place and kill him.
                       -- Mark Twain