2012Feb22

From FilteredPush
Jump to: navigation, search


Etherpad for meeting notes: http://firuta.huh.harvard.edu:9000/FP-2012Feb22

Agenda

  • Discussion of potential grant proposals.
  • NSF Workshop RFP
  • SPNHC
  • AOD
    • Progress on paper, draft circulated
    • Progress on Annotation generation
    • Domain objects supported by Apple Pie.
  • Integration with Morphbank.
  • AnnotationProcessor
    • Graceful handling of open world in ApplePie.
    • UI Design
  • API For Query and Cluster Finding
  • Pending Tech Issues
    • Tech group needs to make a decision on or set a date for decision for the query language for pub-sub for Apple Pie.
    • Tech group needs to decide on or set a date for decision for the scope, composition, and implementation of the "global cache."

Reports

  • Paul
    • Revised BOUML model for network to generate code that builds without errors.
    • Generated code from BOUML model for network and began implementation of Triage and ApplePie job plans.
  • Maureen
    • Worked on the Specify Driver; it can now do retrieval
    • Worked on Morphbank; can now upload our own images to a test instance
    • Worked on Specify collaboration; hammered out a method for running Glassfish in Eclipse
  • David
    • Created code that will use Java reflection to retrieve fields and method return values for use by the annotation generator
    • Created an example xml config document and schema for mapping java fields and java methods to Jena model resources and properties
    • Using jaxb and the reflection code to process the configuration xml and produce the Jena model from the java objects (currently in progress)

Notes

Filtered Push Team Meeting 2012 Feb 22

Present: Jim, Maureen, Paul, David, Bob, James.

Agenda:

  • Discussion of potential grant proposals.

ABI deadline: July 10, 2012 http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=5444

  • NSF Workshop RFP

Bob: Unclear if there is funding, may learn quickly, suggest circulate the idea to a list of potential participants to gauge interest.

  • SPNHC

(1) Demo: Need to sort out hardware issues soon. Two machines, showing both screens at once. Internet access? James to sound out Tim.

(2) Paper/Poster Botany instance.

  • AOD
    • Progress on paper, draft circulated

Bob: Draft circulating. Please comment. Is stuff in the right order? Are things being introduced without being explained?

    • Progress on Annotation generation

David: XML to configure ontology representation in java. Core output in java is JENA model object - can output rdf. Part way through generating rdf concepts from xml configuration. Grouping next week.

Bob: also turtle and N3 as possible outputs. In network, want to transport as rdf xml, might wish to generated human readable documents (e.g. examples from real data).

    • Domain objects supported by Apple Pie.

Bob: Looked again at DwCFP (generated with xslt from DarwinCore doc), has all elements from TDWG DwC Question: Stuff on AppleCore wiki, usage guidance there for recommended elements, doesn't add any semantics. It doesn't look like we need to add any additional terms or semantics within the domain. We can make an apple pie ontology that imports DwCFP and additional desired ontologies (e.g. for provenance, agent identification, grouping, history).

Bob: Is apple core non flat anywhere?

James, yes, identifications, drawing from GBIF extension https://code.google.com/p/applecore/wiki/Identification Complications any time we want to be non-flat.

Maureen: What is a domain object?

Bob: Something which could be the subject or topic of an annotation.

Paul: Yes, also motivation or evidence.

Maureen: What are the things that the database needs to worry about mapping?

Paul: Domain, sensu strictu, darwin core. Broader system accomodates biodiversity domain (domain sensu latu), characters, etc.

Maureen: Looking at morphbank: any morphbank object is domain.

  • Integration with Morphbank.

From Morphbank's online manual (http://www.morphbank.net/About/Manual/annotation.php):

"Note: Currently, only images, specimens and taxon names have annotation options but in future versions, users will be able to annotate any Morphbank object ( i.e. image, specimen, locality, view, publication, annotation, character, etc)."

Discussion. Core domain object for us is darwin core. Starting place for morphbank -FP is specimens and new determinations.

  • AnnotationProcessor
    • Graceful handling of open world in ApplePie.

What is the nature of gracefull handling of Morphbank's wider world of assertions hitting a much more limited set of ApplePie rules and a Specify instance.

Bob: Good to collect open world assertions in annotation store. Purpose of ApplePie rules is to facilitate ingest of data into database in correct place, not filtering out assertions from being made. As long as we can document the vocabularies that are in use in annotations, then people can freely ask questions of the annotation store. (I'd like to query for all annotations that made use of HAO terms).

Bob: Can we ask the annotation store what the namespaces in use are?

Maureen: Lei's UI allow for presentation of incomming in network terms and local in local terms. Straightforward to map non-fitting terms into a remarks field, or let user cut and paste.

James: Need to document and set best practices.

Maureen: Looking at annotation processor, incoming annotation has phenology is flower and database doesn't have a field for phenology.

James: Might throw away, might have manual intervention. Not everyone will have every field in their database. Key is that global knowledge (annotation store) will retain this larger world.

Paul : I like Maureen's cut and paste - user decision.

Consensus: Gracefull handling is throwing away anything that doesn't fit, but presenting it to the annotation processor user and letting them decide to fit it in (e.g. in remarks) where they see fit.

Maureen: Labels for terms?

Paul: List of typical ontology terms in instance, provide internationalization of labels, for those outside, present term name.

    • UI Design

Maureen: Would like to get Morphbank on firuta in a way that is not globally visible.

TODO: Paul and Maureen to discuss requirements, Paul to tighten up firuta access.

For next week:

  • API For Query and Cluster Finding
  • Pending Tech Issues
    • Tech group needs to make a decision on or set a date for decision for the query language for pub-sub for Apple Pie.
    • Tech group needs to decide on or set a date for decision for the scope, composition, and implementation of the "global cache."