Re: AWWSW Telecon Tuesday 2011-05-10

On Fri, May 6, 2011 at 5:18 PM, Jonathan Rees <jar@creativecommons.org> wrote:
> Alan wanted to hear about what we've been up to and where we're going,
> so the agenda I propose is for whoever comes to fill him in.
>
> Jonathan

Some thoughts to get things rolling. Just off the top of my head, so
I've probably left out something important.

AWWSW call this morning - where do we stand - AlanR

Start here:
  http://www.w3.org/2001/tag/awwsw/
  http://www.w3.org/2001/tag/awwsw/2011/status-2011-02.html

Most recent documents:
  http://www.w3.org/2001/tag/awwsw/issue57/latest/
  http://www.w3.org/2001/tag/awwsw/ir/latest/
  http://www.w3.org/2001/tag/awwsw/ir-axioms/
  http://www.w3.org/2001/tag/awwsw/ir-axioms.owl

Web architecture & meaning:
  - normative (protocol & format specs)
  - empirical (descriptive a la linguistics)
  - post hoc
  - wishful thinking (gently prescriptive, a la AWWW)
      'things work better when you do X'
  - prescriptive (put meaning into currently unused message space
      via new RFC or rec)

HTTP is principal embodiment of web architecture:
  Meaning of a message: reasonable mutual expectations of sender and
    receiver.   Contract-like.
  What can we expect:
    - RFCs and Recs followed (usually).  E.g. "authorized" caching
    - unspecified territory not in use (usually)
    - common, but unspecified, practices are followed

Example: What does 302 mean?
  - 2616 or HTTPbis have stories, flawed and probably at variance with
      practice
  - operationally - what do various servers / clients ...
      maybe impossible to generalize (each server has its own rules)

Use cases:
  How to communicate that some 'document' (details TBD) has a
    particular license?  (CC REL)
  How to clarify that a URI is being used to name the IR at that
    URI? (IRW)
  What should Tabulator record as it does HTTP exchanges?

  Add some formalism to AWWW in order to make clearer what's being
  said (and help make sure it makes sense).

or more broadly / vaguely:
  How do we reflect the IETF/AWWW URI namespace semantics in RDF, so
  as to help an RDF application that wants to use its URIs in
  conformance with IETF/AWWW?

Conundrum:
  - you can't impute meaning to an HTTP request or response if it
      wasn't "meant" by the sender (if you can't hold them to that
      expectation)
  - this means spec literalism, and seems to rule out encoding 'best
      practice' like AWWW

Received on Tuesday, 10 May 2011 13:06:37 UTC