From: Miroslav Silovic (miro_at_[hidden])
Date: 2003-07-25 05:23:09
>SMTL aims to efficiently solve queries like:
>a) Which books where written by authors whose last name begins with
> "B" between 1986 and 1994?
>b) How many cities with population above 10,000 are located between
> such and such latitude and such and such longitude?
>c) Which is the closest gas station to this point?
>d) Which are the 10 most similar objects to this complex object?
I'd be very interested in something like this. A few questions:
1. What underlying data structures are used? Are they interesting in
their own right? (in particular, KD trees and R-trees may be useful as
building blocks for interesting applications of their own. For example,
both can be very useful as accelerators for raytracing)
2. Is there any support for more specialised geometric structures? In
particular, BSP trees and Delauney triangulation? If not, are the
structures you've already build easily reusable for this?
3. Have you looked into contacting the author of the relational algebra
boost proposal (this was discussed a few months ago, the archives would
be the good place to look). SMTL would be great for implementing
multi-indexed relational tables.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk