Re: Athena User Interface Project seeks advice

[Maciej Stachowiak <mjs eazel com>]

Havoc Pennington <hp@redhat.com> writes:

> 
> (Also, seriously, if you are trying to develop or deploy software on a
> deadline, a moving target is super-evil and makes it impossible to do
> so. So even though they aren't requiring stock Red Hat, I think it
> would be sensible to decide on their target platform now and stick to
> it until deadlines are met, assuming there are deadlines. Eazel can
> only get away with this for Nautilus because the moving libraries are
> part of the software being developed, and you have direct and
> within-minutes control over said libraries and access to their
> maintainers.)


It's true that Eazel is not your garden variety ISV because what we
are developing is slated to be part of the core GNOME desktop
environment.

But I think anyone else who wants to expend the time and effort could
get a similar level of trust and attentiveness from the maintainers.

And really, the reason we are using the new libraries is not because
we have an easier time getting fixed, but because they provide
functionality we need and reimplementing it ourselves would be
pointless. Dealing with a moving target like Bonobo may inflict some
pain, but it is much less pain than writing our own component model
would have been, and writing our own would not have allowed us to
leverage existing components.

In fact, if anyone else wants to write GNOME software that requires a
component model, I would have no hesitation telling them to use Bonobo
(and improving it if necessary) instead of writing their own component
model. A moving target beats no target.

 - Maciej