Archives for: June 2007

Friday, June 22nd 2007

Permalink 01:10:52 pm, Categories: Meeting summaries

Meeting Summary 18 June 2007

The meeting discussed most of the remaining issues in the Grouping of Resources document. It was resolved that we'll produce a separate document that describes the individual terms (i.e. Classes and properties) in the vocabulary as defined by both the Grouping and main POWDER docs. This will map directly from (and may even be auto-generated from) the actual RDF/OWL data. The discussion that has been taking place on the group's list about redirection was continued. The end result is that by default, if a URI leads to a redirect, that resources will only be a member of the Resource Set if the target URI is itself a member of the set as well. However, this behaviour can be overridden with a new property to be defined. i.e. a Resource Set definition may include a property that says "if redirected, follow the redirection and don't re-process to find out if the new resource is a member of the set." Finally we talked briefly about whether there should be some prioritisation of Resource Set properties. For example, saying that properties that include resources will take priority over those that exclude. There was no consensus on this and more work is being undertaken. The meeting ended with a discussion of the logistics surrounding the forthcoming face to face meeting in Washington.
Phil ARCHER
Permalink 01:08:10 pm, Categories: Meeting summaries

Meeting Summary 11 June 2007

The first order of business was to welcome our new Team Contact, Matt Womer. Known to some group members through the Mobile Web Best Practices Group Matt recently joined W3C from France Telecom's Boston office. The first agenda item was a decision to define a Resource Set as a sub class of OWL Class rather than an RDFS Class. Andrea Perego took the group through the reasons for this but essentially an OWL Class is more closely defined and can allow closed world logic to be performed. The RDF properties currently under discussion for the Resource Sets are things like 'hasAnySchemeFrom' and 'hasAnyHostFrom'. Can these be shortened without losing meaning? Probably yes, but should they? A similar discussion has been held recently in Ubiquitous Web circles and the DDWG which lead to an action to contact Rhys Lewis to find out more! The next discussion concerned whether there should be an RDF property 'hasExactQuery'. We already have queryContains (and queryNotContains) but should we have a more definite version as well? It was resolved that yes we should, however, we should also note that, by their very nature, Web pages generated by resolving URIs with query strings are dynamic and that a Resource Set may not therefore be fully defined even using this method. How should a Resource Set defined by URIs be affected if resolving the URIs leads to a redirection? Should the processor re-calculate whether the resource is or is not a member of the RS or not? The suggestion was made that a 301 9permanent) redirect should be included in the resource set but that other 300 codes should not. This will be discussed further on the mailing list. Finally the group talked about the forthcoming stakeholder meeting in Washington. The signs are very promising that there will be a co-sponsor for this, however the timing needs to be brought forward to the afternoon of Tuesday 10th July. It wasn't possible to resolve the outstanding issues and move to first public working draft of the Grouping of Resources document - maybe next week!
Phil ARCHER

Tuesday, June 12th 2007

Permalink 08:37:02 am, Categories: Meeting summaries

Meeting Summary 4 June 2007

The group is getting close to publishing the first working draft of its Grouping of Resources document and addressed most of the outstanding issues during the meeting. First it was resolved that where regular expressions are used in Resource Set definitions, we will use the Perl 5 syntax. However, we will also note that alternatives are available, in particular, XML REs and and the new syntax under discussion in the X Query WG. Such alternatives may be used in the document if ready in time or in a future iteration of POWDER. The meeting was pleased to note the recent work done by Dan Brickley on the FOAF spec and the discussions under way to create a stable, long term hosting solution for it [1]. Thus the group is confident in using FOAF vocabulary terms in its documents. There was a brief discussion about the stakeholder meeting to be held following the group's next face to face meeting in Washington on July. The venue has now been agreed as AT&T's Innovations Center. Discussion then turned to the Grouping of Resources document. A remaining issue to be decided concerns RS definition by resource property. We want to support a look up system whereby a POWDER module can send a request to a URI and determine whether a candidate resource is or is not an element of the Resource Set by looking at the response. The issue is whether, and how, we should support multiple look up URIs. A proposal will be put to the group during the next week. There is also the issue of a logically inconsistent Resource Set definition - how should these be dealt with? Amazingly, the conjunction and disjunction section that had previously seemed so vexatious now seems to be settled. To do: As above plus: 1. Write normative section giving each RDF property and its constraints 2. Further feedback from the group Aim is to resolve to go to public working draft at next week's meeting. [1] http://lists.foaf-project.org/pipermail/foaf-dev/2007-May/008542.html
Phil ARCHER