IW Meeting 2007 Oct 24
From Inference Web
agenda
- next meeting - Friday Nov 2 11-12:30pm pacific time. Cynthia will send out call in info and reminder.
- progress report
- utep: pml2 provenance data, cynthia to send paulo and allejandro more examples for pml generation by oct 22 what was available in pml 1
- -- provenance data generated by alejandro. DONE
- -- alejandro said it was easier to generate PML 2.0 provenance
- -- cynthia's documentation was good enough for them (and thus may be good enough for others as well)
- -- todo Li review PML validator error message. Li thinks this is a stanford namespace problem.
- Discussion - namespace - which ones do we support?
- we must support inference-web.org. it looks like the ontology list includes:
- http://inference-web.org/2.0/pml-provenance.owl,
- http://inference-web.org/2.0/pml-justification.owl,
- http://inference-web.org/2.0/pml-owl.owl,
- http://inference-web.org/2.0/ds.owl
- Todo - li will update the offending confusing messages
- we must support inference-web.org. it looks like the ontology list includes:
- paulo meet geoff on tptp mapping
- report from paulo includes system db which can be used for points 1-3 data for cynthia to start on the translation work.
- todo paulo will call cynthia to provide the specification and paulo will send a paragraph about each phone conversation. we agreed that we would start on 1-3 of the 9 points in task list
- utep: pml2 provenance data, cynthia to send paulo and allejandro more examples for pml generation by oct 22 what was available in pml 1
1. Register languages 2. Register engines 3. Translate justifications (with unknown rules) 4. Encode queries 5. Encode questions 6. Relate justifications to queries (revisit task 4) 7. Register inference rules 8. Extend justification translation inference rules (revisit task 3)
- alyssa update
- alyssa can generate learned source usage data
- alyssa has database access problems from the wireless network connecting to ksl db server. She is working with martin on this problem.
- Li: iwsearch progress
- Li provided link: http://onto.rpi.edu/iwsearch/
- TODO Paulo will ask a student to start using the search tutorial tomorrow (oct 25). Paulo and the student (leo) will report back on usage issues (knowing the tutorial is rough - constructive suggestions are more than welcome).
- TODO cynthia will include a link from inference-web.org to iw search.
- Li: Iwarchive cache and statistics for IW - proposal at least by next meeting and if easy, also the implementation (week of oct 22)
- Li: James working on wine agent
- alyssa update
- request(update pml2): change domain of hasURL from Information to Document, if a document has property URL, then it is a webpage, its content is stored on the Web.
- [option1] preferably not change this term,
- [option2] provide solid example that why this change is necessary
- transition status
- how servers and files will move
- todo
- Li: investigate mapping foaf to pml2
notes
- IW services
- PML-Analyzer
- validate PML document
- extract only PML data
- generate links to follow
- IW-Crawler
- crawl a URL can follow links
- IW-Archive (use database to maintain URL)
- Cache revisions
- revision metadata
- statistics
- IW-Search
- search pml instances
- PML-Analyzer
actions
- next meeting - Friday Nov 2 11-12:30pm pacific time.
- transition inference-web main site
- James starting on new WineAgent.
pml tools
- [todo Li] pml validator: Li review PML validator error message. Li thinks this is a stanford namespace problem.
- Bad error message "The RDF graph does not contain any PML data." http://iw.cs.utep.edu/earthscience/seismic/2/ORG/ACADEMIASINICAINSTITUTEOFEARTHSCIENCESTAIWAN.owl
- [todo Li] IWSearch:
- Li complete IWSearch tutorial
- http://onto.rpi.edu/iwsearch/ published on inference-web main site
- [need decision] pml namespace
- we support http://inference-web.org/2.0 namespaces
- we also backward support http://inferenceweb.stanford.edu/2004/07/iw.owl and http://inferenceweb.stanford.edu/2004/03/iw.owl
collaboration
- todo paulo will call cynthia to provide the specification and paulo will send a paragraph about each phone conversation. we agreed that we would start on 1-3 of the 9 points in task list
pml2 ontology
- [need decision] eliminate maxcard restriction on hasFormat and hasLangauge because we can see one information may use multiple language ?
- [need decision] distinct Fromat from Language:
- format guides users parsing form rawstring to content string
- language guides users send content string to appropriate language engine
- [need decision] hasURL move from information to Document
- Cynthia will generate initial concern for moving back. e.g. what if both description and content have URL? (Li: they are two different documents, and they are related by hasDescription relation)
- Li believes that hasURL is a property of document.

