From: Joao Abecasis (jpabecasis_at_[hidden])
Date: 2005-03-30 10:44:08
Dan Rosen wrote:
> Hi João,
> Just wanted to send a brief reply. I think you make some good points
> about the diversity of possible node designs, and they're reaffirmed
> by Andreas and David's points as well. I think Andreas has swayed me
> to explore Jeff Mirwaisi's "feature-oriented" tree ideas some more,
> and I think your comments are particularly applicable in that context.
> Have you had a chance to look at the treelib source and docs that
> Andreas posted? I'd be grateful for your opinions on that, since I
> personally have very little experience with that style of programming.
I've downloaded the code but have yet to take an in-depth look. The
approach looks interesting, to say the least, and as to the
possibilities it seems very powerful. However I don't think we've got to
that point of the library design. Well... I haven't, anyway ;-)
IMO, wether, in the end, you go for an all-powerful, adaptable,
extensible tree that performs optimally for any requirement set and
scenario, for many different single-purpose trees, or you pick a mix of
both approaches is not really important, right now.
Perhaps the most important part of a Boost Tree library will be the set
of concepts it endorses. Coming up with a bunch concepts that doesn't
limit the implementation possibilities, I think, is the hard part and it
is where such a Boost Tree library should begin.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk