Re: GEM : Gnome Event Manager

[Owen Taylor <owt1 cornell edu>]

Tom Tromey <tromey@creche.cygnus.com> writes:

> Olivier> It would be interesting to have a central piece of software
> Olivier> through which apps could pipe their important events. It
> Olivier> would allow the user to choose actions associated with
> Olivier> events. This could be done as a new layer over the CORBA
> Olivier> Event Service.
> 
> I agree that we need something like an event service.  I read the
> CORBA Event Service docs today.  It seems really complicated and
> gratuitous, like most of what comes of OMG (IMHO, of course).  Still,
> I don't see any real barriers to implementing the Event Service on top
> of any ORB.

[...]

> How important are the OMG-specified services to Gnome?  I can see us
> spending a lot of time implementing services which, while useful to
> Gnome, will also include a lot of stuff we don't care about (e.g., is
> any Gnome program really going to be hurt if we only have Push event
> channels and not Pull channels?).  On the other hand, these services
> are standardized.
> 
> To me, it isn't clear what the standardization really gains us.  It's
> not like there is a large body of applications that we can drop into
> our framework.  My opinion is that if we're going to be writing all of
> the infrastructure anyway, we might as well write what we really need.
> 
> 
> I suppose one might argue that the OMG designs are complicated because
> they must be, for reasons that aren't immediately apparent.  I wish
> the specs included rationales for the choices they've made.
> 
> 
> I'm curious to know what others think about this.

Well, I don't have much of an answer, but I must say that I have seldom
seen my own thoughts expressed so accurately. I would _love_ to
see rationales for the OMG's baroque designs.

I don't think we should be at all shy about using partial
implementations to start off with, and even, where it seems justified
using our own simpler non-compatible services/facilities. 
Although I think using CORBA has a lot of potential, there is a
big barrier to get over to actually using it, and the more we can
lower that barrier, the better.

Regards,
                                        Owen