[Pellet-users] A lot of descendant queries

[bijan at clarkparsia.com]

On Thu, May 3, 2007 1:50 pm, Alan Ruttenberg <alanruttenberg at gmail.com> said:

> Might merit a special case reasoning technique. I've implemented this
> particular case in the past based on a doubly recursive walk of all
> of the subclass relations and the stated part-of relations, which is
> fast, though I never proved it was complete.
> 
> The same case occurs for any transitive property like part_of - there
> will be others.

There is certainly a bunch of literature on evaluating transitive closure queries in databases. We do want to avoid heavyweight evaluation where a walk will do. It's a little tricky since not every edge is necessarily manifested in any way and we have the possibility of blocked edges (disjoint classes, negated roles).

Cheers,
Bijan.