See also: IRC log
<timbl> The TAG had a diagram to express that
<Stuart> jar: suppose that you have a URI that you dereference and different times requesting different language variants.
<Stuart> ... there is something that needs to in some sense be consistent between the returned representations.
<alanr> http://www.ics.uci.edu/~fielding/pubs/webarch_icse2000.pdf
<Stuart> ... looking for something that 'mandates' such a connection (between representations).
omnigraffle diagram:
http://sw.neurocommons.org/2008/inforesource.png
<Stuart> http://sw.neurocommons.org/2008/inforesource.png
<timbl> tag:Representation
<alanr> http://www.w3.org/DesignIssues/Generic.html
<timbl> Value includes metadata and bits
<Stuart> jar: uses value mean a value that isn't necessarily constrained to be an 'awww:representation' of any particular thing.
<alanr> u:RepresentationInvariant
<Stuart> jar: Fieldings more formal definition of a resource is as a temporal mapping function to sets of equivalent values.
<Stuart> fyi. I think fielding actually says mapping to [sets of equivalent values] or URIs (ie. includes notion of redirection aswell).
<alanr> he writes "any concept that might be the target of a hypertext link" versus "any web page that might be the target of a hypertext link"
<alanr> stuart, yes, including URIs is one of the things that makes it inconsistent
<Stuart> alan... how so?
<alanr> stuart: One inconsistency: Resource = a mapping, versus resource = Any information that can be named can be a resource
<alanr> a document, an image. These are not mappings
<alanr> what about differing laws in us, france re: labeling.
<alanr> If liability is the criterion, then they could have different information
<alanr> so sometimes = same information, sometimes = same liability
<alanr> this can be confusing if not predictable, discoverable, at least for machine interpreters
<Stuart> Well... I think that (on the web) I can see a document as a mapping to values (which happen to encode the state of the resource).
<alanr> stuart, the slogan here is "that's epistemology". One shouldn't confuse how we come to know something from what that thing is.
<timbl> Now looking at which bit?
<Zakim> Stuart, you wanted to note that Fielding (at least orally) take a utilitarian view of resources.
stuart: uri denotes the abstract
resource, cloud is a view
... resource proves its utility by being consistent
alan: it's not clear enough
timbl: commonly, we mean the
abstract document byt the uri
... solid line should be denotes, not abused to denote
... you can't provide a complete definition of IR that will
satisfy everyone
alanr: 'document' had a prior meaning that was different from AWWW sense
timbl: copyright of language variants is held by same entity
alanr: work, expression, performance are different from doccument
timbl: you can't choose one word that will make all communities happy e.g. library vs. business meann different things by 'document'
<alanr> for the OWL document, we have a section introducing OWL to different communities (If you are a database developer.... If you are a java programmer ...) If Document is to be used as an effective term, I think we need to have such a thing.
<timbl> sameWorkAs
jar: can i say sameas for two tag clouds over same AD?
timbl: depends on intent
sameworkas
<alanr> If it can be sameAs to one community, but not use sameAs in a different community, then doesn't this get in the way of sharing world wide?
<Stuart> I think that you could say that they were variantWorks of the GenericWork.
timbl: can't conclude that A is same as B
<Zakim> jar, you wanted to ask about uris for ADs that differ only by tag cloud
<alanr> people use it all the time
timbl: wouldn't use sameas ...
<timbl> bycataloge#awww = <http://www.w3.org/TR/AWWW>.
a.com/foo.html = b.com/bar.html ?
<timbl> <> <>
<Stuart> Xiashou has been using "awww:representation" identifies "awww:resource" with an understanding of "awww:representation" to be an ephemeral thing at a moment... and in that context his "awww:representation" identifies "awww:resource" makes sense.
<timbl> <htp://a.com/foo.html> ex:managedMy [foaf:name "A.com" ]
<alanr> but the value cloud is dependent on what the server wants to do
<alanr> two servers serving the same resource don't have to both give the same representations
<alanr> they only have to give the same representations if they serve the same mime type
http://a.com/foo = http://b.com/bar ?
<Stuart> Alan... what for you makes the resource available at these servers "the same resource"?
timbl: when you refer to it via a
particular service, you mean the AD *as managed by that
service*
... who is responsible for its upkeep?
... for most purposes we don't care
<Stuart> Oh... another thing on Fielding's model is that there is a two step mapping URI->fn(t); fn(t)->[representation | URI]
<timbl> Well, we are allowing { ?x ?p ?o } => { ?y ?p ?o } for app p and o.
<Stuart> which kind of acknowledges that the URI->resource mapping can change.
timbl: there is never a complete semantics
<alanr> Suppose they both are serving a syndicated report of the weather in Oaxaca
jar: but what are ok statements to make?
<alanr> that was answer to Stuart
<alanr> stuart, thanks for the picture :)
<Stuart> I'm looking for the kind of objective test that you would use.
timbl: the pdf version has a length of 3.2 kilobytes
<alanr> Indeed, I would like that too. I think I know how to do that sort of thing, but usually when I try it is considered to formal, or tooo boundary
timbl: value cloud as a set is
not worth talking about
... because it's hypothetical
... different ads each time
<Stuart> I think that the value cloud is an artifact of the way that we (Fielding) model a resource (give it a mathematical formulation).
timbl: misleads people into thinking the resource *is* the set of files
<Zakim> timbl, you wanted to say the abstract resource is what is denoted. Retract my happiness with the document. and to say the abstract resource is what is denoted. Retract my
<timbl> I wonder whether w could have a new picture wth the arrow fixed IMO?
<Stuart> I don't think it is intended to say that all resources are literally mappings of function to time, just that we can choose to model them that way.
alanr: to the extent that we're trying to learn what the resource is, we can't depend on the value cloud
<alanr> "how do we match a specification to an implementation" - that's what we want
<Zakim> alanr, you wanted to bring in the question of how we come to know based on the value cloud, and he dependence of the value cloud on the server
timbl: stuart, can you remember where the edinburgh diagram is?
stuart: it's important that the
uri denote the abstract document
... if people make statements about the representations using
the uri, they're doing the wrong thing
<alanr> blasted phone battery. back in a sec
<Stuart> content-location:
stuart: representations are not the things being referred to ...
<Stuart> Jar... have you taken a peak at Larry Masinter TBD and DURI draft?
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/to/too/ No ScribeNick specified. Guessing ScribeNick: jar286 Inferring Scribes: jar286 WARNING: No "Topic:" lines found. Default Present: TimBL, jar, Stuart, Alan, Alan_Ruttenberg Present: TimBL jar Stuart Alan Alan_Ruttenberg WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 15 Apr 2008 Guessing minutes URL: http://www.w3.org/2008/04/15-awwsw-minutes.html People with action items: WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report[End of scribe.perl diagnostic output]