2010July28

From FilteredPush
Jump to: navigation, search


User: Zhimin Wang | User: Chinua Iloabachie | User: James Macklin | User: Paul J. Morris | User: Maureen Kelly | User: David Lowery | User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.



Reports

  • Chinua: Evaluating RichFaces4, no tab control yet, but has other solutions to our primefaces scaling problems - can scale out UI to display of 1000s of annotations without difficulty. Prototype running PrimeFaces and RichFaces together without problems. Demo is ready.
  • James: BRC proposal with Yale for Imaging herbarium specimens and using several different forms of workflows submitted with FP component, being careful about FP commitments beyond existing prototype.

Brief overview of data flow in BRC proposal, and ways a FP network might fit in, raising the question of who is responsible for writing the code to do the mapping to transport incoming data from the network into a local database schema. Original prototype FP put that as a responsibility of the client. Current view of a bidirectional mapper is moving the coding into the responsibility of the FP network (with the actual mapping of field to field being done by client users but with a FP mapping component.

  • Zhimin: Has been investigating the structure of IPT and the computer science literature on schema mapping and on data exchange (and the broader data integration literature). Algorithms are published for doing the desired sorts of complex mappings, but no open source code bases that are relevant to us. To evaluate use of IPT we need to better understand how complex the required transformations will be. IPT does generate a configuration file for the mapping in a particular instance. Nightly build is now working, with a pair of cron jobs running the build and doing the reporting. Report is at: http://gentoo.cs.umb.edu/svn/FP/nightly_build_report.txt There is a standard called web service choreography language for coordinating several different web services, might allow web service from each node to be composed by an outsider. Architecture MS is in Bob's hands for editing. Quering MS to be done next, literature review in progress.
  • Maureen: Researching project management techniques. Sent out email summarizing roles. Joel on software (blog) provides lots of good pointers to how to develop successful software, key repeated message is to understand who you are writing the software for and what the software does.


  • Paul: Reported on current state of annotation ontology. Discussion: Not clear what the consequences of assertions are when annotations are expressed in domain ontology concepts. We need to produce some more instances where there are annotations of concepts expressed in owl, as well as the current instance that illustrates an opaque object.

Deliverables

Proposed list of deliverables for documentation of the prototype to be completed by the end of August (Maureen to overview and check for internal consistency):

1) Concept of Operations: Description of the system from a user's point of view (James).

2) A set of use cases [Paul].

3) A set of requirements [Paul].

4) Description of prototype components, including documentation of prototype system design (including both component and class diagrams) [Zhimin, with Chinua [WebClient] and Maureen[specify]).

5) Documentation of Message and Annotation schemas and ontologies [Bob, Zhimin].

6) Evaluation of prototype technologies (Hadoop, ActiveMQ, Specify, IPT, JiBX, JSF), with discussion of alternatives [Zhimin, David, Chinua].

Reports to be developed at: Prototype Reports