Re: OpenOffice and GNU GLUE (was: Is GLUE active?)

[Markus Fleck]

"David A.Cobb" wrote:
> You may have noticed that there's been a lot of discussion about groupware
> and collaboration on  <address@hidden> lately.  I know some of it
> spilled over onto the glue list, and its beind more or less duplicated on
> the Evolution list. [...] So, shall I turn the stampede in your direction?

Thanks for the offer.

I should note that a stampede of GNUs is nothing to start lightly,
especially as GNU GLUE is still rather a couple of tiny flowers
which could be trampled down easily before they're even blooming...

I've been looking at cross-platform GUI options for the last couple
of weeks. The two main contenders are OpenOffice's GSL/VCL and the
FOX GUI toolkit library.

I have a couple of problem with OpenOffice at this time:

- It's a HUGE piece of software, both source and binaries (cf. Mozilla)

- Documentation, especially of internal APIs and architecture, is
  sparse and will probably remain so for some time (cf. Mozilla)

- OpenOffice is considering to replace or augment GSL/VCL with GDK/GTK
  and UNO with Bonobo; I believe that that is a very bad idea and might
  put the future of the whole OpenOffice project at risk (cf. Mozilla)

- It's extremely hard to understand OpenOffice internals unless you
  have lots of time to spend on it (such as when you're paid to do so
  full-time :-); this issue will get even more problematic if OpenOffice's
  internals should become an unstable or moving target, i.e. if the projects
  decides that substantial parts will be rewritten (cf. Mozilla)

- As a side note, integrating with OpenOffice's object model might make
  GNU GLUE vulnerable to StarBasic and JavaScript macro viruses - unnecessarily
  so. Scripting languages should only go so far, and (IMHO) a groupware system
  shouldn't be able to execute JavaScript/BASIC scripts from untrusted sources
  at all (cf. Mozilla's rather problematic, IMHO, XUL<->JavaScript integration)

That said, I'd like to propose the following "roadmap":

FOR OPENOFFICE:

1. The OpenOffice project should try to replace the missing scheduling and
   mail/news components; I probably won't be able to help with that, but I'm
   confident that some people will start to work on this sooner or later.

2. OpenOffice should implement basic WebDAV support (read/write/lock/folders).

FOR GNU GLUE:

3. GNU GLUE should develop a consistent groupware framework with some reference
   GUI client implementation, probably using FOX as the GUI toolkit, and
   design its internal protocols and data structures such that the 
implementation
   could eventually be ported to the OpenOffice infrastructure.

4. Some fine day, maybe next year or in the year after that, the GNU GLUE
   reference client could be ported to and integrated with OpenOffice.
   Even "deep" integration (such as for real-time distributed editing)
   might be feasible. This presumes that a) GNU GLUE progresses nicely
   in the meantime and b) OpenOffice doesn't become an "Officezilla"
   nightmare.

I would guess that adding more "groupware" functionality than 1. or 2. to
OpenOffice might make the OpenOffice project unmaintainable. IMHO, it's
crucial that OpenOffice starts with small steps, such as reintegrating the
"missing parts" and writing documentation, code samples and programming
tutorials, so as to make it easier for external contributors to participate.

So unless your personal interests are in experimental collaborative groupware,
everybody should concentrate their efforts on helping OpenOffice get off the
ground. OpenOffice is an extremely important project, and will need to attract
a lot of very bright hackers in order to survive.

Yours,
Markus.

GNU GLUE Groupware Project - http://glue.sourceforge.net - http://glue.swiki.net