See also: IRC log
<jar> Draft of my versioning memo: http://sw.neurocommons.org/tmp/versioning.txt
FRBR document: http://www.ifla.org/VII/s13/frbr/frbr.pdf
Alan: An Item is an instance of a Manifestation class.
jar: All the exemplars of a Manifestation have the same bits.
(looking at diagram on page 13)
jar: I would draw another arc to "PhsycalBitString".
Alan: Two different printings with different covers would be different Manifestations.
<jar> no, to bit string. the item is the physical thing
Stuart: What if it were a play?
Alan: You have an idea for a story, Romeo and Juliet, independent of language. Then an expression is a performance of it, such as the author writing a version of it or some people playing it out. The manifestation is something like the recording of it -- video or sound recording, or a book. Then the Item is a CD or a physical book.
jar: What if it's PDF?
Alan: With a lithograph, there are slight differences between each physical book, but not with a PDF.
David: With a PDF, there doesnt' seem much difference between Manifestation and Item.
<scribe> ACTION: Alan to prepare material on generically dependent continuum for next meeting [recorded in http://www.w3.org/2008/05/27-awwsw-minutes.html#action01]
<trackbot-ng> Created ACTION-5 - Prepare material on generically dependent continuant for next meeting [on Alan Ruttenberg - due 2008-06-03].
<jar> generically dependent continuant -- from BFO
Alan: IEO = Information Entity Ontology
<alanr_> digital entity and non realizable information entity
<alanr_> BFO = basic formal ontology
<jar> that was DENRIE
Alan: What about FRBR?
jar: People make names or catalog
numbers for things and then they talk about them. If you look
at FRBR, they dont' say whether they are classes or instances,
but we talk about them adn there's somem world view, and I'm
trying to come up w a method that allows any such world view to
align w web semantics.
... That's the acid test for a theory of a web semantics: can it integrate someone else's view? Suppose Work is a class and Tim's IR is a class. Is Work a subclass of IR? You'd want to be able to answer that to know. Web semantics isn't a semantics unless you're on the road to being able to answer that.
<alanr_> dropped. calling back in
Stuart: I wonder if there are things that are IR that are Works, but perhaps not all IRs are Works.
David: I think the question is the wrong way around.
<jar> dbooth: semantic web arch is a model. you have a choice as to how you cast things into it.
<jar> dbooth: not a question of is or isn't it an information resource, but how one chooses to make the correspondence
<jar> dbooth: the owner of the uri gets to choose ...
<jar> dbooth: question of work vs. nonwork is independent of IR vs. non-IR
<jar> jar: what is not an information resource?
<Stuart> I think that things that have mass are not information resources.
<jar> dbooth: an IR is a function
<jar> jar: IR's don't have mass. maybe we can agree on that.
<jar> jar: boothian IRs don't have mass. good.
<Stuart> FWIW: I think Roy's model is a model of the operation of Web Infrastructure in respect of an Information Resource - ie. the mechanism by which an IR is inspected by a GET operation and reported on in the form of a awww:representation.
<Stuart> ie. modeling an IR as a function from time to sets of awww:representations isn't saying that an IR *is* such a function, but that a way of modeling a view of a resource as presented by web infrastructure in the form of awww:representations is as such a function.
<jar> dbooth: work vs. IR (function) would be a pun
<Stuart> jar... I wonder if it would be useful to run through the Journal Publication example in 1) FRBR and 2) as web resources
<alanr_> worms r us
<jar> dbooth: punning is inevitable
<alanr_> ir = work, representation = expression?
<alanr_> journal article = expression
<alanr_> bound in a magazine = manifestation (each is an item)
dbooth: Punning is undesirable but inevitable, because it is always possible to come up with a finer identity distinction.
<jar> dbooth: undesirable but inevitable, because it's always possible, and sometimes necessary, to make finer distinctions
<Stuart> jar... what is the thing in what you are discussing that motivates a possible change in URI?
<jar> scenario: a FRBR adherent defines U to be a FRBR Work.
<jar> that was 1.
<jar> 2. they publish the work (or manifestation of it) on the web at U, with 200 response
<jar> 3. apply for a $1M grant which is contingent on adhering to web architecture
<jar> question: do they have to mint new URIs?
<jar> stuart: answer might depend on what exactly they think U identifies
<jar> stuart: the Work is abstract, and doesn't have representations
Stuart: I think they don't necessarily have to change their URIs. It may depend on whether they think they are identifying a Work or Manifestation. They're clean if they thinnk they're publishing a URI of a Manifestation of a Work.
<jar> stuart: it's not a concrete document as such...
<jar> alanr: compare to Tim's bylaws example
Alan: Tim say a generic document is like the by-laws of an org, and a representation is like a PDF.
<jar> stuart: generic documents - you end up with a pdf or html - you also get an alternate URI that is more specific
<jar> alanr: is moby dick not a generic document?
(jar needs to leave)
Dbooth: My answer is partially written up here: http://dbooth.org/2007/splitting/#httpRange-14
<alanr_> I.5a: Implement FRBR concepts to present related works hierarchically, pulling together all records related to a particular work (e.g., Moby Dick), diverse expressions of that work (e.g., translations into German, Japanese and other languages), different versions of the same basic text (e.g., the Modern Library Classics vs. Penguin editions), and particular items (a copy of Moby Dick on the shelf).
<alanr_> from http://www.frbr.org/2006/01/24/rethink
<jar> jar will probably say: no one except web arch geeks *want* to talk about boothian function/IRs. so no classes of things that anyone wants to talk about will ever intersection function/IR. ergo no one will ever want to use 200-responding URIs to name anything that they want to talk about.
<Stuart> The intention of webarchitecture is that a given URI consistently refer to the same thing.
<Stuart> The harder part is saying with precision what thing that is - which I think is Pat's principle point.
dbooth: The basic idea in http://dbooth.org/2007/splitting/ is to show how to deal with situations in which a URI is minted and it later turns out that that URI is ambiguous. For example, a URI for AKT is minted as an instance, and it later turns out that there are three distinct things AKT1, AKT2 and AKT3. Architecturally, this is no different than when Mark Baker uses the same URI to denote himself and his web site. Hence, "the use of a URI to directly d
enote both an information resource and a non-information resource should be viewed as a violation of good practice, but not a violation of Web architecture".
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/FRBR/FRBR document/ Succeeded: s/Entity/Item/ Succeeded: s/continuum/continuant/ Succeeded: s/basic formal ontology/BFO = basic formal ontology/ Succeeded: s/have/have to/ Succeeded: s/say/saying/ Succeeded: s/think/thing/ Succeeded: s/I thing/I think/ Succeeded: s/what think/what thing/ No ScribeNick specified. Guessing ScribeNick: dbooth Inferring Scribes: dbooth Default Present: Jonathan_Rees, DBooth, Alan Present: Stuart David Alan Jonathan Got date from IRC log name: 27 May 2008 Guessing minutes URL: http://www.w3.org/2008/05/27-awwsw-minutes.html People with action items: alan[End of scribe.perl diagnostic output]