[BioSQL-l] latest changes

[Aaron J Mackey]

On Wed, 2 Apr 2003, Hilmar Lapp wrote:

> The last one (term_path_id in term_path) is somewhat artifactual, in
> many respects motivated by the bioperl object model where
> Ontology::Path is its own object, not just an association (like
> bioentry_relationship and seqfeature_relationship). Since there really
> isn't a fundamental reason why Ontology::Path should be the only object
> which is not identifiable by primary key, I think it is OK to obviate
> the need for lots of special case code by just adding a PK column to
> term_path. If you don't have an object corresponding to term_path, the
> PK doesn't need to bother you.

I'm fine with this; in fact, as I've said before, I'm in favor of every
table having a table_id surrogate primary key (with other, natural PK's
enforced via unique constraints).  I believe that makes for a more
flexible schema.

-Aaron

-- 
 Aaron J Mackey
 Pearson Laboratory
 University of Virginia
 (434) 924-2821
 amackey at virginia.edu