See also: IRC log
<scribe> Scribenick: mhausenblas
<scribe> ACTION: Michael to check missing pointers from AWWSW to minutes and add if necessary [recorded in http://www.w3.org/2009/03/10-awwsw-minutes.html#action01]
<trackbot> Created ACTION-15 - Check missing pointers from AWWSW to minutes and add if necessary [on Michael Hausenblas - due 2009-03-17].
<dbooth> Jonathan's diagram: http://www.w3.org/2001/tag/awwsw/jar-diagram-3.pdf
dbooth: had some difficulties
understanding the diagram
... reason might be that arrows are used a bit unsual
<dbooth> Michael: No huge difficulty understanding jar's diagram, but unsure of the overall goal. What are we really after?
Jonathan: a starting point it is ;)
Michael: would like to clarify the goal, first, if possible
dbooth: from NS to response from service?
jar: timbl and stuart wanted to differentiate between HTTP entity and representation
dbooth: HTTP message and response from service different boxes?
jar: probably the same, but from HTTP in the two different ways (communication vs REST)
<dbooth> michael: I read Roy's ACM article, but we should be careful what part we adopt as a basis -- that whole thing regarding ont and epist thing. Worth summarizing what's in there. I could summarize it.
<dbooth> jar: It's not open access -- requires permission to access.
<scribe> ACTION: Michael to summarise Roy's article on REST [recorded in http://www.w3.org/2009/03/10-awwsw-minutes.html#action02]
<trackbot> Created ACTION-16 - Summarise Roy's article on REST [on Michael Hausenblas - due 2009-03-17].
<dbooth> michael: We usually, with ont, try to come up with instances. Are we ready to do that?
<dbooth> michael: as long as we're only talking at the metalevel we'll get into endless loops.
jar: I can populate it
... example is the Generic URI note itself (the HTML page)
<dbooth> michael: Until that is done, it may be hard to communicate. Instances may help.
dbooth: how about an example with a cookie
jar: what's the difference to UA?
dbooth: a good second example might be the 'my bank account', yes dbooth
<dbooth> michael: If you book a flight, it depends on the country what content you see, etc.
<dbooth> alan: That's exactly why that idea of a time varying info thing doesn't make sense. Missing link to external world.
Alan: link to the external world is missing in that diagram
jar: lot of different things to say with 200, right?
dbooth: let's choose a canonical example
Alan: where my questions re the relations answered
jar: yes see
... I think what bothers Alan is the formality
... for example, the Moby Dick example (abstract, not bound to the real word thing)
scribe doesn't understand Alan (to noisy) - sorry
Alan: time-varying concept doesn't make sense
Alan: for most parts of the
diagram they are grounded
... it is not about modelling, it is about capturing, common understanding, definitions, etc
<dbooth> michael: Why don't we model obvious things like URIs, status codes, etc.? I believe everything should be grounded, but we then hit a meta layer where we define it.
jar: let's relabel the box
dbooth: let's rename it to TimBLsGenericResource
Michael volunteers to set up the ontology part
Michael: let's put jar's current diagram as an RDF-S schema into a Neologism instance (run through Pellet online then)
<dbooth> dbooth: Proposed example instance of "TimBLsGenericResource": a web page that is occasionally updated and is served in multiple languages and encodings.
<jar> please: URIs contain accession numbers - the labels are human readable but labile - URIs stable
<dbooth> dbooth: +1 to michael's offer :)
<scribe> ACTION: Michael to code jar's diagram in RDF-S [recorded in http://www.w3.org/2009/03/10-awwsw-minutes.html#action03]
<trackbot> Created ACTION-17 - Code jar's diagram in RDF-S [on Michael Hausenblas - due 2009-03-17].
jar: canonical examples
... FTRR - bank account
<jar> what about the AWWSWVocabulary
Alan: the cookies violate the principles, not everything is conveyed via the REST modelling
<lori> jar: can you send me the graffle?
<dbooth> dbooth: Canonical example of ftrr: dereferencing http://citibank.example/balance yields dbooth's balance if dbooth's login cookie is sent in the request.
Michael: Roy's ACM paper explicitly mentions cookies breaking the REST model
<alanr> a bank account is not a representation
<alanr> update of "loose talk sinks ships" -> "loose talk crashes mars rovers"
<jar> a bank account page (like main account page) *can have* many representations... one for each account... and "constructed" or rendered on demand, not stored
jar: how does he REST model handle cookies, etc.
<alanr> for frbr, closest to expression is the box time varying information thing
<alanr> for frbr item just means instances
jar: a REST resource has a membership function
dbooth: what about time?
Michael: what about a human in the picture minting a URI and assigning meaning
<Stuart> Bummer.... I goofed with the hour change.
jar: the three models are TimBLsGenericResource, REST resource, HTTP response
<jar> 3 points of triangle: generic resource, REST, http response source
jar: next meeting on 31 March
This is scribe.perl Revision: 1.133 of Date: 2008/01/18 18:48:51 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/XXX/REST/ Succeeded: s/mybank/citibank/ Succeeded: s/*has*/*can have*/ Found ScribeNick: mhausenblas Inferring Scribes: mhausenblas Default Present: DBooth, mhausenblas, +1.617.538.aaaa, Alan_Ruttenberg, Jonathan_Rees, Stuart Present: DBooth mhausenblas +1.617.538.aaaa Alan_Ruttenberg Jonathan_Rees Stuart Agenda: http://lists.w3.org/Archives/Public/public-awwsw/2009Mar/0009.html Got date from IRC log name: 10 Mar 2009 Guessing minutes URL: http://www.w3.org/2009/03/10-awwsw-minutes.html People with action items: michael[End of scribe.perl diagnostic output]