Re: awwsw status report for TAG F2F

Comments not inline below.

Thanks for raising this, Tim. While agreement on things that are
either good or correct or inconsequential is a good thing, and
agreement on nose-following would be a very good thing to the extent
it were one of those, I don't want the recommending to be part of our
charter, for two reasons. First, it makes AWWSW's chances of failure
much greater, as it would be difficult to find a more philosophically
diverse group of people that us. Second, I think that as a matter of
process the recommending (to the community) has to be left up either
to the TAG or to some future WG. (I suppose we could recommend
something to the TAG.) Third, I think we should try to do one thing,
namely an ontology, and do it well. Going beyond analysis to
recommendation would threaten the quality of the analysis effort, I
believe.

(If we could simply explain our differences to everyone's
satisfaction, without worrying about who is right, that would be a
tremendous accomplishment.)

How about the following compromise?

* To investigate and explain one or more possible RDF-yielding
nose-following methods.

I recognize that ontologies can act as propaganda, and I'm happy with that.

-Jonathan

On Thu, Jun 18, 2009 at 7:30 AM, Tim Berners-Lee<timbl@w3.org> wrote:
> Nice summary. One comment inline.
>
> On 2009-06 -16, at 10:48, Jonathan Rees wrote:
>
>> I just whipped this up.  The deadline for TAG F2F prep materials is
>> today, so if you can comment in the next 3 hours that would be great.
>> Otherwise you'll just have to live with my disclaimer. If necessary I
>> can provide corrections and further information at the F2F.
>> - Jonathan
>>
>> AWWSW Task Force - status report for TAG
>>
>> Jonathan Rees, June 2006
>>
>> This is my own report, and has not been reviewed by the group.
>>
>> Regular members:
>>
>>  * Tim Berners-Lee
>>  * Jonathan Rees
>>  * Michael Hausenblas  - joined Feb 2009
>>  * Stuart Williams
>>  * David Booth
>>
>> Irregular members:
>>
>>  * Alan Ruttenberg
>>  * Harry Halpin
>>
>> Contributing by email occasionally:
>>
>>  * Pat Hayes
>>  * Henry Thompson
>>  * Noah Mendelsohn
>>
>>
>> Home page: http://esw.w3.org/topic/AwwswHome
>> Mail archive: http://lists.w3.org/Archives/Public/public-awwsw/
>>
>> We have been holding biweekly telecons.
>>
>> Generally speaking, we are trying to put the relationship(s) between
>> HTTP and RDF on a more rigorous footing.  Among the questions falling
>> under this objective are:
>>
>>  * If an HTTP interaction (such as a 200 response or redirection)
>>   can be said to say something about a resource, how might one
>>   capture that meaning in RDF?
>>
>>  * If we know something about a resource that impinges on HTTP
>>   behavior, such as representation stability over time, how might
>>   one express that in RDF?
>>
>>  * How might one express in RDF (e.g. via class membership) some of
>>   the differences between entities and practices that respect web
>>   architecture, and those that don't?
>>
>>  * Is there any basis for recommending any particular RDF-yielding
>>   nose-following method over another?
>
> This question is not IMO on the table in that the whole purpose of W3C
> is to develop standards for interoperability. The questions of algorithms
> for RDF-yielding nose-following are I think things whose standardisation is
> critical
> and where the TAG's job includes making sure the larger scale picture is
> clear
> and not just the individual granules of spec.
>
> * To investigate and recommend RDF-yielding nose-following algorithms.
>
>> I do my best to steer the group away from unanswerable questions like
>> "what is an information resource" and "how should the use of GET+200
>> be restricted", but I think everyone has present in their mind the
>> goal of being able to say something to rationally inform the
>> httpRange-14 debate.
>>
>> The group has not yet produced any consensus artifacts, but is hoping
>> to publish a draft ontology by the end of 2009.
>>
>> The following have been discussed:
>>
>> == Ontological diversity ==
>>
>> We have compared the myriad ontologies and models of HTTP:
>>
>>  * HTTP spec
>>  * REST informal description (Roy)
>>  * REST formal description
>>  * Generic resource "genont" (Tim)
>>  * AWWW (TAG)
>>  * Boothianism (David Booth)
>>  * Resources that you can access (Pat Hayes)
>>  * IRW (Harry Halpin)
>>  * HTTP-in-RDF  http://www.w3.org/TR/HTTP-in-RDF/
>>
>> Many of these are placed next to one another in this diagram:
>>
>>  http://w3.org/2001/tag/awwsw/jar-diagram-5.pdf
>>
>> == Generic resources ==
>>
>> We spent a while trying to nail down what Tim means by "generic
>> resource" or "information resource".  The results are collected here:
>>
>>  http://esw.w3.org/topic/AwwswGenericResource
>>  http://esw.w3.org/topic/AwwswGenOntDiagrams
>>  http://lists.w3.org/Archives/Public/public-awwsw/2009May/0006.html
>>
>> == HTTP semantics ==
>>
>> My proposed semantics of HTTP, based on a suggestion of Pat's, is
>> here:
>>
>>  http://esw.w3.org/topic/AwwswRfc2616
>>  http://w3.org/2001/tag/awwsw/jar-diagram-6.pdf
>>
>> This is very new and has not been discussed by the group yet.  It only
>> assumes that the two parties in an HTTP exchange are obeying the HTTP
>> protocol specification.  There is no assumption that they know
>> anything about web architecture.
>>
>> == Recurring problems ==
>>
>> Ground rules.  Different participants have very different ways of
>> doing ontology design, with Harry and Alan perhaps representing
>> opposite ends of some spectrum.  Solution: Do not talk about it.
>>
>> Slippery terminology.  "X is a representation."  "No, Y is a
>> representation."  Solution: Focus on definitions (or some other
>> anchor), not words.  Punish unqalified use of terms when choice of
>> definition is not clear.
>>
>> Confusion between syntactic things vs. events (e.g. response, entity,
>> representation).  Solution: Always articulate which is meant.
>>
>> == Earlier work ==
>>
>> Errata for the httpRange-14 resolution
>>
>>  * http://esw.w3.org/topic/ErrataHttpRange14
>>
>> Vocabulary survey (stale)
>>
>>  * http://esw.w3.org/topic/AwwswVocabulary
>>
>> Questions and use cases (needs to be revisited)
>>
>>  * http://esw.w3.org/topic/AwwswQuestions
>>
>> Some N3 rules around HTTP (needs to be revisited later)
>>
>>  * http://esw.w3.org/topic/AwwswDboothsRules
>
>

Received on Thursday, 18 June 2009 12:55:22 UTC