I think this is a great idea, if only for getting some people testing these user contributed features.  If these are never tested then how will they ever get approval for contribution into the main svn trunk?  It is tough for normal users to search through lists posts to find a patch someone has contributed for the functionality they are looking for.  If there was an open repository where users could submit unofficial patches, I think it would make it much easier.  On the opposite side of the argument if submitted patches were reviewed regularly and feedback was provided for any bugs or documentation fixes, then the open repository would probably not be needed.  I am not sure which approach is better in the long run, but I am definitely in agreement with Andreas that the work that has been done in the bindings is great and I hope the community keeps expanding as well as the functionality supported by the bindings.

Jesse

On Sun, Apr 6, 2008 at 12:09 PM, Andreas Klöckner <lists@informa.tiker.net> wrote:
On Samstag 05 April 2008, Andreas Klöckner wrote:
> Hi all,
>
> I've rolled a new bindings snapshot, available from [snip]

Somebody contacted me off-list and suggested that what I said might be
perceived as accusatory, and that my intentions weren't entirely clear. So
let me clarify:

First off, I'm not placing any blame for dropping patches. I realize that
there may have been some issues that ultimately prevented their inclusion.
What I'm advocating (and providing) is a proving ground for contributions to
the bindings that has a lower barrier to entry than the official sandbox svn.

It seems that most contributions to the bindings are of the drive-by kind, but
I believe that that doesn't make them worthless. I'll just simply collect
them, and whatever is deemed fit for inclusion in the official tree, can go
there with no problem. I don't mean to maintain an incompatible branch, or
imply any criticism of the current maintainership. I just believe there's a
niche for a faster-moving, include-now-ask-questions-later version of the
bindings. Sort of like -mm vs -linus in the kernel, if you wish.

In summary: If you, as a user, are willing to accept fewer guarantees in
exchange for more functionality and more fixes (and also, potentially, more
breakage), then you are more than welcome to use my tree. If not, then don't.

Andreas

_______________________________________________
ublas mailing list
ublas@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/ublas