Re: Opendoc (was Re: spell checking)

[Miroslav Silovic <silovic zesoi fer hr>]

Martijn van Beers <sauron@il.fontys.nl> writes:

> I think it would be a good thing (tm) if Linux got something like
> OLE (i.e. Opendoc), no matter how opposed most of you are to the
> 'bloat' of CORBA, office suites and the like. If GNOME wants to be
> the user-friendly desktop system I hope it is going to be, we need
> to look at the users wishes, not just our own. And most users are
> used to, expect and demand a good office suite.

Actually there are at least 3 possibilities (probably more):

	1) OpenDoc

The problem is that OpenDoc might not be as free as we'd like. It's a
good idea to check with Apple first. Its one drawback is that it
requires SOM beside CORBA. Don't ask me what SOM is, I only know it's
'System Object Model'. Another problem with OpenDoc is that it's big,
and dictates much about the application architecture.

	2) Fresco

Fresco is free software, under X licence. At one point, it was
distributed with X in WorkInProgress (first release of X11R6 had
it). Then it died from lack of funding, and because it came at the
wrong time (3-4 years ago, OpenDoc seemed viable and OLE seemed
unbeatable, and there weren't any free CORBA implementations to work
with). Fresco is really neat in design, and gtk implements a few of
its concepts (for example, the use of boxes for layout, or arbitrary
widget embedding). Fresco used CORBA and allowed free embedding of
applications. We could probably move many Fresco concepts on top of
Gtk (although not all of them, as Fresco allows some crazy things like
arbitrarily rotated widgets - buttons slanted by 20 degrees, for
instance).

Lately the activity has somewhat resumed, at a rather low key.  Their
website is at http://www.iuk.tu-harburg.de/fresco and I think it's at
least worth the look for ideas - and maybe contacting the developers
for possible cooperation.

	3) Look into GNUStep/BeOS

I think both have some sort of application embedding, much simpler
than OLE but therefore more workable. And interoperability with
GNUStep, at least, seems very desirable to me.

	4) roll our own

Since there is only one standard that's currently actually being used,
and that one is absolutely out for us (I'm talking about OLE), there's
no problem whatsoever with being incompatible. That's IMHO.

-- 
I refuse to use .sig