Hilmar writes (to Arlin):
BTW did you see the Nat Biotech article on the OBO Foundry (http://dx.doi.org/10.1038/nbt1346)? The NCBO is also preparing a guidance document for current and prospective ontology contributors (the latest revision is attached). You or at least Enrico (assuming that he is spearheading the ontology?) may want to be subscribed to obo-discuss, the mailing list where these things are discussed (Julie may be on that list already, though).
Linking between CDAO (or glossary) and nexml
subject of a telecon 1/25/08 with Rutger, Enrico, and Arlin
- added dozens of terms to glossary in November and December
- not much else since then
- some work with Julie on transformations, initial ontology
- started to get familiar with glossary
- applied for NESCent support to work on CDAO this spring
- set up nexml.org site
- met with Kansas group to come up with plan for NeXML and publication
- support in 5 languages
- Perl via Bio::Phylo, work by Rutger
- Python via standalone, BioPython-compatible module (Jeet)
- Java via
- XML Beans
- generic SAX class libs (i.e., should work with BioJava, Mesquite)
- C++ via XSD similar to Java XML Beans
- support by 5 application
- support in 5 languages
Arlin: need to rationalize nexml web site and NESCent wiki to avoid overlaps and out-of-date content.
Rutger: division will be
- nexml.org has code, auto-docs, schema
- NESCent wiki has project planning, design philosophy
description of the problem
Rutger took us to nexml.org.
- We looked at the Complex Type called "TreeFloatEdge".
- the documentation has a description of this entity
- the description is actually an "Annotation" in the schema
- "Annotations" could be used to hold links to the glossary or CDAO
- glossary needs to be formalized to be used
- but glossary is informal logically, need ontology
- could use glossary as initial relation-less version of ontology
- not all one-to-one links will be available, right?
- that is, some nexml entities do not map directly to CDAO entities
- but the descriptions include primitives that could map to CDAO
- also the OWL language allows statements to define an entity
- for instance, if tree is defined but not TreeIntegerEdge, the OWL language allows for statements that restrict tree such that the edges have only integer values for lengths
Conclusion of the above is that we could put the links into the annotations in two ways:
- annotation is OWL statement defining exact equivalent of nexml entity in CDAO
- annotation is natural language description with some terms that are exact equivalents of CDAO concepts
The second method is easier to begin with (while CDAO is incomplete), and perhaps its easier to implement and update, but its a much weaker (less expressive) link between nexml and CDAO.
Enrico volunteered to come up with suggestions for the two main issues:
- how to represent the glossary in a formal way, e.g., as the initial version of CDAO
- how to make links from nexml to the glossary-CDAO in a way that can be updated
Then Enrico will send this to us for comments, and we'll go from there.
Also-- because we need to be aware of, and take advantage of, opportunities to promote the project and to coordinate with other projects that are related--, I will look into PhyloCode and see if we have any interest in the Halifax meeting this summer.