[CMake-Promote] OpenSceneGraph

[Brandon Van Every]

On 12/21/05, Brandon Van Every <bvanevery at gmail.com> wrote:
>
>
> I can and will inquire.  However, I can anticipate the outcome.  Negative
> results may be a good lab for what issues we need to address, though.
>
> Meanwhile, if people can find leads on *current* conversations about
> CMake, that is much more useful.
>

More recent posts indicate they may be "vulnerable."
http://openscenegraph.net/pipermail/osg-users/2005-December/060142.html
It seems their releases are taking them too damn long.  A unified build
system is perceived as one way to alleviate this.  I'm not sure if ctest /
dart coordination would also help them.  Seems possible; I don't know much
about those tools, so someone else should eyeball them from that
perspective.  OSG looks like it has a very large user community, so it may
be worth a lot of mindshare points to pocket them.  Which, aaah, looks like
a big chunk of work since they have all sorts of dependencies under OSG
itself.

I wish we could go after the core libraries that everyone uses, and work
from dependencies upwards.  I see that as unlikely to actually work though.
Core libraries, being core, are heavy duty tested under GNU Autoconf all the
time.  So there's little incentive to "fix" their builds.  Far more likely
that we'll sway people with big lumbering projects that desperately need a
unified cross-platform solution.  Which means, we're going to face this
problem of somebody needing to do lotsa lotsa work over and over again.  We
should think about how we might present ourselves to such projects and
ameliorate such work.  Especially since it's the large projects that have
the most mindshare attached to them.


Cheers,
Brandon Van Every
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/cmake-promote/attachments/20051222/52fdd989/attachment-0001.html>