Re: reviewing the charter of awwsw

Hi Jonathan,

Many thanks for the overview, it's really helped as I wasn't sure quite 
where to begin.

Of the various points mentioned there are a few which I'm deeply 
interested in (or should I say have spent a lot of hours focussing on), 
and others where I don't feel I understand the issues well enough to 
comment yet.

I'd like to start by summarising my current thinking on information 
resources, the dual use of URIs, and then a few commments on the 303 
pattern (from both a network and semantic perspective) then hopefully 
lead on to some of the other topics like promises of instability, and 
HTTP Semantics, the meaning of each status code.

Thus, I'll strike whilst the iron is hot, and summarise my current 
thinking on the aforementioned, hopefully as proper introduction to the 
group, and possibly as a starting point for a fresh round of 
conversation. I'll start now and send to the list when ready.

Best & thanks again,

Nathan

Jonathan Rees wrote:
> Feeling a bit exposed by the addition of a new person, I just took a
> few minutes to look at the genesis of our charter, which was recorded
> here:
> 
> http://www.w3.org/2001/tag/2007/11/05-afternoon-minutes
> 
> Here are some of the topics that were discussed at that meeting:
> 
>   'information resource'
>   how to make the semantic web work reliably
>   what URIs to use in HCLS context
>   RDF & URIs in published papers
>   URI insurance policies
>     e.g. fallbacks when links are found to be broken
>   promises of stability (TBL suggested 'fixed resource')
>   warnings of instability ('this is a demo')
>   dual use of URI in denoting and providing access - the "magic of webarch"
>   trust earned by info: and LSID
>   unreliability of 303
> 
> Somehow the precipitate of all this was to convene a series of calls
> "on HTTP semantics". I'm not sure this wording accurately describes
> what those who signed up for it had in mind - there's no record
> linking what transpired in the meeting to the action to form the
> group.  It may have been scribed this way because of Dan Connolly's
> observation that something was the "same thing as Tim's call to do a
> group to formalize HTTP" (for which I don't have a reference).
> 
> So where do we stand on the above issues?
>   - Little tangible progress on 'information resource' but I haven't given up.
>   - Henry T. and I have been working on reliable URIs as TAG ISSUE-50.
> This is very hard and outside the scope of AWWSW.
>   - What URIs to use for HCLS - this question ended up being outside
> AWWSW - e.g. the "shared names" coalition took Dan C's recommendation
> seriously; still somewhat hostage to the reliable URIs question.
>   - Fallbacks - as far as I know the only movement on this is the
> statement in the GBIF identifiers report (which I helped write), which
> indeed is what Dan C would describe as taking parts of URI space by
> eminent domain.
>   - warnings - not much demand yet, but demand might be stimulated by
> any new persistence solution
>   - dual use of URI  - related to 'information resource' - this has
> ended up taking most of my attention here in AWWSW
>   - info: and LSID - this question has faded from view as LOD blooms -
> info: seems to be moribund, and LSID has only a few new adopters (in
> biodiviersity informatics) - however URNs generally continue to have a
> following, and the reliability of http: is still an open question
>   - unreliability of 303 - LOD is exerting a lot of standardization
> pressure here.  Alan R is probably still bothered by the lack of any
> expectation setting.  In my mind it's less of an issue than I thought
> it was at the time (I didn't know no non-LOD 303s were deployed, and I
> didn't know how ready HTTPbis would be to endorse GET/303)
> 
> Questions like "how do you figure out what a URI means" didn't come up.
> 
> TimBL presciently said "A danger is failing to focus the group
> sufficiently narrowly.  Risk is that the group tries to do
> everything." He was referring to the HCLS URI note I think, but that's
> way too broad and goes way beyond 'HTTP semantics'. The need to limit
> scope may explain why it ended up as "HTTP semantics," putting
> reliability and nose-following out of scope for us, and why I've
> insisted on hammering on the "dual use" issue, which is the only
> remaining intersection between the discussion at the meeting and the
> heading "HTTP semantics".
> 
> So I have now refreshed my understanding of why this is the topic that
> I've been pushing on so hard here.
> 
> In spite of this focus I have found it very difficult to untangle the
> issues. For me to understand "dual use" with any depth it seems I have
> to pick apart all the core web specs (3986, 2616, AWWW) as well as RDF
> semantics; on top of this I need to understand document semantics
> (still an open problem in library science) and all sorts of
> philosophical nastiness I'm unqualified to process.
> 
> The rest of our work on HTTP is still relevant, but relatively easier,
> so I'm not too worried about it. If we had a customer for it that
> would help move it along faster.
> 
> Reliability and nose-following are clearly important. Maybe they
> should have their own task forces.
> 
> Jonathan
> 
> 
>

Received on Friday, 14 January 2011 23:56:30 UTC