Semantic Web Interest Group IRC Scratchpad

Welcome to the Semantic Web Interest Group scratchpad generated automatically from discussions on IRC at irc.freenode.net port 6667 channel #swig by the chump bot, instructions in the chump user manual. Please use UTF-8 charset on IRC and pastebin for code or data more than 10 lines long.

Nearby: IRC logs (Latest) | semantic-web list | W3C Wiki (Recent changes) | delicious swigbot

last updated at 2004-11-10 11:46
 
 
 
libby: by Ikki Ohmukai, Hideaki Takeda, Masahiro Hamasaki, Kosuke Numa, and Shin Adachi
libby: can't find a preprint
libby: this is their site
libby: includes: Egocentric search, RNA/Glucose, foaf trackback and something else scribe missed
libby: "lightweight publishing"
libby: ah, danbri's had a look at their system: see his chump entry
 
libby: full paper preprint (pdf)
libby: by Kendall Clark, Bijan Parsia, Bryan Thompson, Brad Bebee
 
DanC: 19:00Z, 14:00 ET, 11:00PT on Monday November 15th 2004.
 
libby: by Jonathan Hayes, Claudio Gutierrez
libby: presentation at ISWC data semantics track
libby: full paper pdf preprint
libby: Jonanthan Hayes yesterday, codepicted with me:
libby: Libby Miller, Jonathan Hayes at ISWC 2004 photographed by Jeen
jeen: jonathan's approach looks very similar to the way in which sesame's main memory store indexes graphs
jeen: that approach was chosen for compactness: unique object for each rdf term
 
libby: Ramanathan Guha, Richard Fikes, Rob McCool
libby: presentation at ISWC data semantics track
libby: TAP website - data scraped from many sites; also text extraction of news items
DanC: I wonder if I need to read this or if it's pretty much the same as his previous work on contexts.
DanC: nearby: ContextLogic.lsl, my larch-noodling on Contexts: A Formalization and Some Applications from Guha at stanford
libby: "various contextual dependencies, e.g. by time/space "clinton is president" many more..."
libby: "in databases the contextualization is there but implicit and hidden"
libby: "so they're pretty sure it's here to stay"
DanC: reminds me also of when I saw Lenat talk about contexts at KT2001
libby: "scraped data is contextualized, because based on language for humans"
libby: "their proposal: SW triples are true in a context"
DanC: this is at least similar to, if not the same as, the N3 construct: <foo.rdf>.log:semantics log:includes { :clinton :is :President }
DanC: documents either are or have contexts. I'm not sure which.
DanC: briefly read it. very interesting point on the spectrum between minimal/elegant and quick-and-dirty.
DanC: I think it's not quite right in saying that N3 is just like RDF reification, but that's not his fault; we haven't written N3 up well enough yet.
DanC: here's hoping for time to try expressing their rules for importsFrom and such in N3 rules.
 
libby: by Mark van Assem, Maarten R. Menken, Guus Schreiber, Jan Wielemaker, and Bob Wielinga
libby: pdf preprint of the full paper
libby: "start with a syntactic conversion, then a semantic conversion using owl primatives"
libby: [interesting...some of the guidelines look pretty similar to the ideas we used for converting icalendar to RDF]
libby: "Interpretations of the meaning of information in the original source (i.e., meaning that cannot be traced back to the original source or documentation) should be approached with caution, as wrong interpretations result in inconsistent and/or inaccurate conversions."
libby: "Instead of developing a new schema (i.e., thesaurus metamodel), one can also use an existing thesaurus schema, such as the SKOS" (scribe note: yay!)
libby: the website including data and schemas
libby: alan rector: be wary of mesh - the paths are not unique identifiers
libby: presentation at ISWC data semantics track
 
RDF Calendar to iCalendar format using XSLT
DanC: back in May, libby gave a heads up about draft-hare-xcalendar-01
DanC: but the discussion goes back at least as far as Dec 2003: Syntactic profile of RDFiCal?
DanC: my input on the syntactic profile thread came a bit slowly, but it was basically thumbs-up
logger: See discussion
DanC: hmm... <CAP:query>SELECT VEVENT FROM VAGENDA WHERE UID = 'abcd12345'</CAP:query>
DanC: I got started on toIcal.xsl. v1.5 passes test/spec01-conf3.rdf
DanC: I considered starting with MK's code but I'd like to run the whole problem thru my head and update RdfCalendarDocumentation as I go
DanC: toIcal.xsl doesn't actually document the syntactic profile; it just assumes it
DanC: we could use (1) a converter that groks RDF at the graph level and spits out the right XML syntax, and (2) a .rng or .xsd schema
 
to RDF/A or not to RDF/A
DanC: Best Practices WG says yes
DanC: Bjorn says no
logger: See discussion
DanC: esp sbp's pointer to augmeta
 
DanC: by Kenneth Baclawski1, Christopher J. Matheus, Mieczyslaw M. Kokar1, Jerzy Letkowski and Paul A. Kogut
libby: can't find a paper
libby: presentation at ISWC 2004 ontologies session
libby: idea is to develop an onotology to describe inconsistencies in ontologies
libby: ConsVISor tool
libby: related url, Vistology
libby: jeremy carroll has a list of criticisms which I'm not following....
 
libby: presentation at ISWC 2004 ontologies session
libby: preprint of the full paper
libby: "great that lots of large stanmdard ontologies are becoming available, however they are huge, difficult to understand, and 'unreasonble'"
libby: e.g. foundational model of anatomy: you probably just want partof it, e.g. the part related to breast cancer. So how to specifying an 'ontology view' like this?"
libby: "e.g. 1) as a query for a specific set of concepts with particular characteristics or (2) by specifying traveral rules - this traversal view is the topic of the talk"
libby: "e.g. start here, and take this relationship this far and this relationship this far. can be transitive closure for a property or everything related. worked quite well in the breast cancer case"
libby: implemented as a protege plugin
 
libby: preprint of the full paper (pdf)
libby: presentation at ISWC ontologies track
libby: "comparing string similarity of labels; comparing objects; and also set similarities"
libby: "combining these similarities measures - e.g. by averaging - and then measuing against a threashold"
libby: "reduce the number of comparisons, e.g. by randomising, or just choosing the labels with same first 3 letters; once a mapping is found just check nodes local to that one, using ontology structure"
libby: "did it work? tested two travel ontologies created by students, 500 entities. the students also did some mappings between the two. 2 different versions, one with similar labels, one without. QOM approach with similar labels went quite well, quality was reasonable and faster than other approaches."
libby: "with different labels, the QOM accuracy was much lower but improved over time and was much faster than other approaches"
libby: [looks pretty neat. wonder if it really is just label comparison or whether I missed something about set comparison. hm, think I did. read the paper everyone!]
libby: "useful for ontology merging, query routing"
libby: q: why is opnthefly mapping important - isn't offline enough? a: e.g. p2p query routing, ask query in your own ontology.
libby: q: another possible usecase - clinical records from a different hospital for emergency cases
 
 
Created by the Daily Chump bot. Hosted by PlanetRDF.