AWWSW -- 16 Feb 2010

<mhausenblas> heya

hi michael

<mhausenblas> ouch, again your telecon bridge, hm?

sure

i'll get it for you

danc, w'er using my bridge i gave you in private IRC

danc, shall we switch to zakim?

<jar> hello... no luck with zakim just now

<DanC> symptoms?

<jar> i sent email to amy yesterday, too late i guess

<DanC> passcode?

<jar> 'that code is not valid'

<jar> i'll try again

<DanC> I just did an instant telcon

ok, we/ll switch to zakim.

<mhausenblas> ACTIONS?

<DanC> trackbot, action-1?

Semantics of HTTP

<trackbot> Sorry, DanC, I don't understand 'trackbot, action-1?'. Please refer to http://www.w3.org/2005/06/tracker/irc for help

<DanC> action-1?

<trackbot> ACTION-1 -- Stuart Williams to send us a link -- due 2007-12-04 -- CLOSED

<trackbot> http://www.w3.org/2007/awwsw/tracker/actions/1

<mhausenblas> http://www.w3.org/2001/tag/awwsw/http-semantics-report.html

<DanC> (I can show you the logic I used in http://www.w3.org/2001/tag/dj9/story.html )

danc: What problem are we solving?

jar: Expressing semantics of http interaction in rdf.

<DanC> . http://www.w3.org/2001/tag/dj9/httpspeech.html

danc: http can be described to any level of detail

dbooth: yes, but we don't yet have agreement on any levels of detail. we need to choose both a level of detail and reach agreement.

danc: let's look at http://www.w3.org/2001/tag/dj9/httpspeech.html

<DanC> _:some_connection ht:requests (_:some_request).

danc: in this model, _:some_request has time in it, so two different requests might send the same bytes.

<DanC> information(message, resource, hunk_of_data)

danc: you could also use a 3-ary relationship.
... I also did something on caching, not included in this page

<DanC> (caching thing is in http://www.w3.org/XML/9711theory/HTTP )

jar: to use timbl's notion of FixedResource, you need to express time.

danc: to deal with time in the way I've approached it, you'd say that for a FixedResource the max cardinality of :information property is 1.

<Zakim> mhausenblas, you wanted to ask DanC about status code vs method interdepencies

<mhausenblas> http://tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-08#section-9.9

michael: can we use the "via" header?

<DanC> <http://foo> log:uri "http://foo"

michael; (looking at why you use log:uri )

jar: in the other paper i want to focus on what the server is saying about the resource.

<DanC> (other classes in tim's thing? pointer?)

dbooth: 301 (Moved Permanently) would seem to have to deal with time.

<DanC> caching thing... http://www.w3.org/XML/9711theory/HTTP

jar: one of my goals was to show how little there is in the semantics of http interactions, as a way to motivate more layers.

<DanC> I think an information property (which is time-oblivious) and an information2 (which needs a better name and is time/message sensitive)

jar: e.g., the httpRange-14 rule is murky.

danc: i think httpRange-14 is a social convention that is followed, and agents shoudl follow this convention.

<jar> Alan R is urging the OBO ontologies never to use 200 for any URI in any ontology, because the specs are screwed up, in his view

<jar> i.e. always use 303 (or #, but Alan doesn't like # for other reasons)

<mhausenblas> Michael strongly disagrees with Alan's views, btw ;)

<jar> i.e. OBO should follow the TAG's advice, but finds doing so inconsistent with 200, since it's unclear what an "information resource" is or isn't

<DanC> . ISSUE: definitions of HTTP semantics (near "information resource" / 200 responses) sufficient to satisfy OBO designers, who are avoiding 200

<mhausenblas> +!

<mhausenblas> +1

<DanC> ISSUE: definitions of HTTP semantics (near "information resource" / 200 responses) sufficient to satisfy OBO designers, who are avoiding 200

<trackbot> Created ISSUE-3 - Definitions of HTTP semantics (near "information resource" / 200 responses) sufficient to satisfy OBO designers, who are avoiding 200 ; please complete additional details at http://www.w3.org/2007/awwsw/tracker/issues/3/edit .

<DanC> hs:information s:subPropertyOf c:says.

<mhausenblas> ISSUE-3?

<trackbot> ISSUE-3 -- Definitions of HTTP semantics (near "information resource" / 200 responses) sufficient to satisfy OBO designers, who are avoiding 200 -- RAISED

<trackbot> http://www.w3.org/2007/awwsw/tracker/issues/3

<jar> informationResource subclassof Principal

<DanC> "Indirection needs Redirection" http://www.w3.org/2001/tag/dj9/badmeta.html

<jar> server says (resource says X) does not imply server says X

jar: I view awww:InformationResource as a subclass of Principal.

danc: regarding http://neurocommons.org/page/HTTP_Semantics how about using messages instead of time? Time is partially ordered in http.
... There's a :happensBefore relation between messages.
... In caching protocols, the second request is causally related to the first response, as evidenced in the cache validation header. E.g., you get back a picture with etag 100. Then you send another request includeing etag 100.

<DanC> http://www.w3.org/XML/9711theory/HTTP

<DanC> ∧ mp < mq2

danc: but i didn't handle VARY in that caching stuff.

<DanC> represents(mp2, i, cbody)

<jar> look at IAO information artifact ontology

<jar> http://code.google.com/p/information-artifact-ontology/

<DanC> ACTION: DanC look at IAO and propose HTTP semantics sufficient to address their concerns [recorded in http://www.w3.org/2010/02/16-awwsw-minutes.html#action01]

<trackbot> Created ACTION-35 - Look at IAO and propose HTTP semantics sufficient to address their concerns [on Dan Connolly - due 2010-02-23].

<mhausenblas> title for http://www.w3.org/XML/9711theory/HTTP

<DanC> "Specifying Web Architecture with Larch" -- http://www.w3.org/XML/9711theory/

<DanC> Dan Connolly

<DanC> 9th International World Wide Web Conference

<DanC> Amsterdam

<DanC> May 2000

<DanC> ACTION-23 due + 2 weeks

<trackbot> ACTION-23 Review IRW and HTTP Ont OWL. due date now + 2 weeks

<DanC> looking at http://www.w3.org/2007/awwsw/tracker/agenda?duebefore=2010-03-03

<DanC> JAR: yes, next meeting 2 weeks time

<DanC> (2 March)