[Biojava-dev] release 1.4 timeline

Michael Heuer heuermh at acm.org
Wed Mar 24 15:06:53 EST 2004


On Wed, 24 Mar 2004, Matthew Pocock wrote:

> Hi,
>
> What's left to add before we make a 1.4 release? It's been a while since
> we did a release and 1.3 is starting to look old. Here are my requests:
>
> 1) Make biosql work
>   * pass unit tests
>   * support ontologies

What about a simple implementation of biosql ontologies, in other words,
just do a data-blob version of the biosql schema?

ie.

Ontology
{
  String getName();
  String getDescription();
  Set<Term> getTerms();
  Set<TermRelationship> getRelationships();
}

Term
{
  String getName();
  String getDefinition();
  String getIdentifier();
  boolean isObsolete();
  Ontology getOntology();
  Set<Dbxref> getDbxrefs();
  Set<TermRelationship> getRelationships(); // where this is the subject
}

TermRelationship
{
  Term getSubject();
  Term getPredicate();
  Term getObject();
  Ontology getOntology();
}

Dbxref
{
  ...
}

> 2) Make ontology support work 'well enough'
>   * Having written a number of reasoning engines, none of which will be
> maintainable by anybody other than me, I propose we just use Jena and
> largely dump our code - http://www.hpl.hp.com/semweb/

Having made some attempt at digging around the bowels of the ontology
package, I think I might agree with this.  You're often working way over
my head.  :)

For what it's worth, I do like my own pet project at

> http://vocabulary.sf.net

which has, among other things, distinction between terms used as concepts
(subject and object in term relationship) and those used as relations
(predicate term), and a representation of the assignment of a term to
something else, supported by evidence (e.g. the GO annotation files).

It does not have any support for reasoning, however.

> 3) Documentation (Javadoc) fixes - look at the doccheck reports and act
> accordingly
>
> What else needs doing? Who will volunteer for parts of these?

Personally, I'd be much more interested to start work on a "biojava2",
whatever that may mean, but I'll try to fill in where needed.

   michael



More information about the biojava-dev mailing list