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

[George <jirka 5z com>]

On Fri, Dec 01, 2000 at 09:53:26AM -0800, Darin Adler wrote:
> The gnome-vfs API is only necessary for Nautilus because we built on top of
> it at Miguel's suggestion. Are you saying that gnome-vfs is now considered a
> technical dead end? If so, Miguel and Ettore essentially led us into a trap,
> because we had planned our project without gnome-vfs and they strongly
> suggested that we use it, saying that all of GNOME would be using it for
> GNOME 2.0.

[ note that this is sort of a reply to the whole thread ]

Gnome-vfs is not only neccessary for nautilus.  It is neccessary for 2.0
GNOME.  While monickers are cool.  They are NOT the solution to absolutely
everything.  Here we have to think of the users.  Not of perfect technical
ideas.  We are not building Hurd or Berlin, which will be perfect, but never
usable.  We are building something which is to be usable NOW.

1) gnome-vfs is the more straight forward way to access files in a virtual
   fashion, something which all applications need to do.  Not all
   applications are going to use monickers overnight.  Some applications
   might not use monickers at all for reading and writing files.  We must
   ensure that apps can see/read/write all the files that nautilus does.

2) gnome-vfs is used NOW in nautilus, gnome-vfs works now.  No matter what 
   technical reasons might be against it, it's here and no it doesn't suck.
   My mom doesn't care that she'll be able to use a beautifully object
   oriented desktop 20 years from now.  These are the types of users we need
   to work towards.

Miguel: It was your one premise in your talks, that microsoft is winning
because they are willing to accept non perfect solutions and build on top of
them.  So let's do that.  We DO want to beat microsoft.  Using a non-perfect
(and note that many people (including me) think it is quite a cool solution)
solution like gnome-vfs is part of the way there.

So let's stop bickering and look at what we can do to beat microsoft:

1) Release 1.4 as soon as possible.  This means freezing things and accepting
   dependencies such as gnome-vfs or whatever.

2) Work towards things we don't do yet, rather then reinventing things we can
   already do.  We'll have time to redo it all on our next desktop project
   when GNOME has got 90% of the market.

3) Let's NOT care about dependencies.  Dependencies happen.  I don't see a
   problem here.  It's the same as requiring gnome-libs for bonobo, or
   some of the other fairly arbitrary dependencies.  I mean oh well, let's
   move on.

As for GNOME 2.0: 

1) gnome-vfs support is being added in gnome-libs, and I'm
   very intent on this being this way.  Why?  reasons outlined above. 
   Why not do everything with some "perfect" solution?  We have too many other
   problems that are deeper and worse that need to be solved, not to mention
   we'd have to agree on the perfect solution.

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.  1.4 unfortunately doesn't add the "big step" we need, and 2.0
   does.  We need to get 2.0 out SOON.  If anyone wishes to work on the
   perfect solution to this problem I'd consider it a 2.2 or 3.0 (or
   whatever) work.

George

-- 
George <jirka 5z com>
   Computers are useless. They can only give you answers.
                       -- Pablo Picasso