Future Data Exchange Standard/Teleconferences

Teleconference, Feb 3, 2009

Mark, Peter, Jeet (Kansas); Rutger; Hilmar, Jim (NESCent); Enrico; Arlin (recording)

1. NeXML C++ parser status
   • Mark tried auto parser generated from Code Synthesis, experienced difficulties incorporating this into NCL, dedcided to give this up, though someone else could pick it up; did not recall in detail what were the reasons-- possibly some problems with compiling and with validating id links
   • (Hilmar) GPL a barrier to commercial implementations such as PAUP*, prefer L-GPL
   • (Mark) ok to relicense NCL with L-GPL
2. NeXML Java parser status (and Mesquite integration)
   • (Rutger) implementation uses config file in Mesquite folder
     • parses some test files
     • but can't parse Phenex files (http://phenoscape.svn.sourceforge.net/viewvc/phenoscape/trunk/data/phenex-files/)
3. Ontology-mediated dictionary attachments
   • (Jim) Phenex-- part of Phenoscape project; effort to translate data sets associated with Zebrafish project
     • each OTU annotated with dict references to OBO id, OBO specimens, OBO collections
     • each state annotated with phenotype terms
   • (Rutger) Sawsdl (http://www.w3.org/2002/ws/sawsdl/), in schema have model ref attribute of URI to ontology term; provides uniform syntax for referring to ontology terms
     • (Jim) example problem: on the fly extension of ontology, e.g., bony process on hyoid bone
     • (Enrico) can't do that with sawsdl
     • (Rutger) important for hackathon to come up with way to add ontology-mediated meta-data, not just using term directly, but extending
     • (Enrico) on-the-fly annotations are instances
     • (Jim) don't want to create new terms; (Peter) want to avoid too many terms in ontology
4. Character partitions; Referencing character columns in a matrix: separating ordering of characters from the referencing strategy
   • (Jeet) referencing character columns by position rather than element index; should we allow both?
     • (Hilmar) strongly in favor of element ids
   • (Jeet) character partition concepts; currently have no facility to refer to char partitions or blocks
     • (Rutger) already have facility in schema for defining sets of characters
5. Identification of the "niche" / use-case scenarios for NeXML 1.0 (given its more lightweight scope, i.e. sans models etc.)
   • (Jeet) where do we see 1.0 being used?
   • (Hilmar) declare 1.0 before or after hackathon? Let this be driven by the hackathon.
6. Target state of development for "NeXML 1.0" feature set (and maybe a freeze/release date)?
   • expect that this is going to come together with hackathon

We adjourned at 2:45 Eastern time.