SWD WG

 

 

<RalphS> Previous: 2006-10-10 http://www.w3.org/2006/10/10-swd-minutes.html

Admin

Guus: PROPOSE to accept minutes of 10 oct telecon ...

tomb: SECOND

guus: so RESOLVED.

tomb: one person is always taking notes into irc, then captured into a file, then end of meeting, IRC is also command line for set of programs, one of which takes the irc log ...

and turns it into minutes ...

every week somebody is scribe, includes taking raw output of program which makes draft meeting report and cleaning it up ...

tomb: good to have meeting reports that are both quotable text to reply in email and as HTML files to browse ...

please consider using a mailing program that handles quoting well ...

guus: conventions for scribing actions etc. ...

ralph: ben adida joined me in my office ...

notational conventions, use IRC to scribe so corrections can be made on the way ...

<RalphS> Ralph: irc notation is important -- e.g. "Ralph: irc notation is important" means the scribe heard Ralph say those words

many tools for producing minutes easier ...

notational conventions ... when citing a person speaking, put that person's name or irc nick followed by colon ...

then text quoting them ... note that in other irc communities a colon means you're addressing the person not quoting them ...

<RalphS> Ralph: "... signals continuation line"

ralph: alistair should be using dots at the beginning of the line ...
... script that makes the minutes will also recognise the dots at the end of the line ...
... we'll find out ...

<RalphS> ACTION: Ralph consider a tutorial on scribe conventions [recorded in http://www.w3.org/2006/10/17-swd-minutes.html#action01]

ralph: actions and decisions ...
... can do "ACTION:" in uppercase or lowercase
... at the end of the meeting when RRSAgent is excused he reports the list of actions
... one other IRC convention is, when I typed "ACTION: ..." the bot responded in a slightly different format, depending on the IRC client
... what I see is asterix then "RRSAgent ..."
... this is an "offline" or "out-of-band" comment
... style, those things do not get recorded on the irc log, do not show up in the minutes, use this for anything that doesn't want to go in the permanent record
... use your judgment about what is "off the record"
... equivalent to styles used e.g. in swig

<RalphS> Ralph: "/me is what my irc client uses for the other style of off-the-record"

ralph: those two styles are the same as far as the meeting record is concerned

<Zakim> TomB, you wanted to ask how to get a URI such as http://www.w3.org/2006/10/10-swd-minutes.html#action01

ralph: note that tomb used "q+" to raise his hand
... note ralph ack tom
... RRS agent builds those URLs (for actions), RRSAgent initially builds the URL that points to the raw IRC log, timestamped ...
... action URL is produced by the minutes reformatting script ...
... can ask for the minutes to be redrafted, gets queued in the background
... can now look at the -minutes.html file.

tomb: those action URLs get generated at the end of the telecon?

ralph: get generated by the draft minutes request.

daniel: how does scribe get designated? all minutes get recorded at the server side, what does scribe have to do after?

ralph: don't have to put "scribe" in your nick".

<RalphS> scribenick: alistair_scribe

ralph: script will use a heuristic to determine who was probably the scribe, if scribe changes midstream can change scribe explicitly ...
... that text in the irc record tells script which lines coming from scribe
... other question, the IRC log is recorded on the W3C servers, the draft minutes script will produce an HTML file which always needs a little bit of hand editing
... e.g. scroll to today's record, see there is diagnostic output, in there will be diagnostics pointing to places that might need some attention
... once you've dealt with issues, delete diagnostics. If you've followed the conventions for scribing then the amount of cleanup is minimal.

daniel: where do you go to clean up the minutes?

<RalphS> today's minutes

ralph: when you ask for the minutes to be generated, you'll get the URI ...
... constructed according to a regular pattern.
... because minutes into year/month directory on W3C, cannot give write access to WG participants, so we work around this ...

ralph: e.g. when scribe has finished cleaning up the HTML file send it to me, I'll put it up.
... i'll find a link to some info about this ...

<TomB> http://www.w3.org/2005/rules/wg/wiki/Telecons

<RalphS> Chris Welty's collection of how W3C meetings run

<RalphS> Art of Consensus

ralph: art of consensus started out as hints for WG charis, been expanded to include other stuff ...
... see especially section "collected wisdom and advice"

<TomB> http://www.w3.org/Guide/#mtg-advice

<scribe> ACTION: ralph to link to chris' telecon notes from SWD homepage [recorded in http://www.w3.org/2006/10/17-swd-minutes.html#action02]

guus: customary to announce yourself when beginning talking
... especially at start of new groups.
... tom & I will also propose scheme for scribing.
... going back to the agenda, one action in the minutes for ralph

<TomB> ACTION: Tom and Guus to propose a scheme for scribing [DONE] [recorded in http://www.w3.org/2006/10/17-swd-minutes.html#action03]

<RalphS> other potential WG meetings 22 Jan

guus: straw poll last week on f2f, good timing wrt planning

guus: f2f meetings typically where we decide to release key documents

guus: timing would be right to release use cases & requirements for SKOS

<RalphS> ACTION: [DONE] Ralph communicate to January meeting planners our desire to keep the option for f2f open for a couple more weeks [recorded in http://www.w3.org/2006/10/10-swd-minutes.html#action01]

guus: any new views on this? week of 22 january, for 2 day meeting either at meginning or end of meeting

ben: if I'm in town I will attend

ralph: monday and tuesday best for me

guus: same holds for me

rachel: same for me

guus: assume for the moment 22 23 january, need to decide on this within two weeks

SKOS REQUIREMENTS

<TomB> scribenick: TomB_scribe

<TomB> Alistair:

<RalphS> scribenick: TomB

Alistair: On requirements. Today posted a long email to list on thoughts on requirements for SKOS. Based on my paper. Suggest we be careful not to bite off more than can chew.

<RalphS> SKOS requirements [Alistair 2006-10-17]

AlistairScope ourselves tightly by limiting use cases to two Patterns:
... 1) one controlled vocab used to "index" some objects, then same vocab used to query index
... focus in terms of functionality: all to do with retrieval, specifically retrieval using Controlled Vocabularies (CVs) ("Pattern A")
... 2) more than one CV in play. When different CVs used in metadata and in query, mapping is required ("Pattern B").
... in DC-2006 paper, sketched cases, broad-brush. What features draw out?
... So far no negative feedback.
... If we decide to accept this focus on retrieval, then difficulty: to describe the functionality in the abstract (not getting caught up in User Interface issues)
... Wrote thesis on this - see posting.
... Different types of visualizations of CV for humans - which is necessary for SKOS to automatically generate?
... Emphasize Semantic Web flavor of use cases -- merging data from multiple sources.
... emphasize distributed aspect, data merging, in use cases.

Guus: Feedback regarding different use cases (different vocabularies)?

Bernard: Could not access papers - without tasks, not clear.

Alistair: "indexing" is the term that this community uses -- constructing metadata about objects and using controlled vocabulary for constructing metadata.
... "indexing": associating metadata with objects.
... CV index associates one or more terms from the vocabulary.
... an "index" is "some metadata involving controlled vocabularies".
...Example: library catalog with Dewey code.

Bernard: why is the metadata important? Why not just, intrisically, the vocabulary?

Alistair: taxonomy or CV needs to be designed with a function in mind.
... CV for one purpose will be different from another, e.g. Wordnet - structure of vocabulary will have different features because applied differently.
... by focusing on functionality we want to support, focus on features we want to capture in the language itself.
... if our representation is sufficient for the functionality.
... Without focus on functionality, could be hard to understand when we have satisfied requirements.

Guus: People here using vocabularies for purposes different from indexing?

<RalphS> http://www.webster-dictionary.org/definition/etiology

Daniel: shared repositories - desktop applications - medical ontologies. Problem is that content is not stabilized, so use cases not established.

Guus: Daniel, is what Alistair is saying consistent?

RalphS: Similar reaction to Bernard. Alistair mentioned patterns A and B. As SWD moves forward, examples of deployed systems - specific customers in mind, helpful to cite early on.

Guus: Alistair, give pointers to existing systems.

<scribe> ACTION: Alistair give pointers to deployed SKOS systems. [recorded in http://www.w3.org/2006/10/17-swd-minutes.html#action06]

RalphS: you have abstract cases - continue in abstract vein?

Alistair: in my DC-2006 paper, a hypothetical use case, concretely illustrating Pattern A - also in dissertation

<Zakim> seanb, you wanted to ask whether there are clear examples that are *out of scope* of Patterns A and B

Alistair: my posting today has relevant links.

SeanB: whether there are clear examples definitely out of scope of these patterns?
... helpful if there are examples of things we do _not_ want to address.

Alistair: Yes, e.g., Roget's Thesaurus (or thesaurus in MS-Word).
... glossaries and dictionaries...
... "terminologies" (confusingly called) - sets of lexical data for supporting translators
... if we accept my suggested scope they would be outside _our_ requirements (not to say it couldn't be done)

Rachel: Are these patterns SKOS-specific?

Alistair: these are application scenarios, within which SKOS can be applied in order to communicate shared data between software components involved.
... trying to illustrate use cases at high level...
... then saying "these are the types we want to target and support"...

Rachel: need clarification of context

Antoine: pattern A is a traditional library use case
... where the corpus is indexed by a controlled vocabulary in order to facilitate retrieval.

Rachel: so these patterns can be applied to more than SKOS?

Alistair: yes, these patterns can be applied to other organizations and services
... we're trying to enable functionality in a distributed setting; i.e. Semantic Web, not necessarily provide new functionality.

Guus: On mailing list this week, let's exchange whether these are useful scoping restrictions (or too strict)
should plan for requiremtns WD to come out in january ...

Guus: need to decide who will coordinate this draft
... who will be willing to volunteer
... normally would not ask main designed to coordinate the requirements document
... talk about next week

RDF in XHTML

ben: task force in last WG (SWBPD) focus on rdf in html .
... continued work in summer, just implementation and demo

ben: now getting back into working on RDF/A syntax and primer docs

<RalphS> update on RDFa progress from Ben

ben: update with lessons learned, (lack of) difficulties in implementation
... RDF/A is a way to express RDF triples in XHTML, using mostly attributes that already exist in XHTML, apply a bit more liberally to other elements
... main construct is the "rel" attribute in the "a" element
... e.g. if you link to creative commons license, and add a rel atttribute, you are effectively creating an RDF triple
... so how easily can we get authors to embed rdf triples in html

<RalphS> RDF in XHTML Taskforce (old)

ben: RDF/A primer and syntax docs, working hard to update with latest changes

<RalphS> RDFa Primer editors' draft

<RalphS> RDFa Syntax editors' draft

ben:RDF/A primer is good intro, short, examples, if it doesn't make sense give us some feedback
... happy to have more contributors, ??? from IBM joined in summer

guus: issue, do we continue with the idea of task forces, groups working seperately and reporting back to the main group
... is your proposal to continue that way? (ralph has doubts)

ben: ralph's concern regarding fragmentation makes sense, but there are some fairly detailed rdf/a issues to discuss that may not be of interest to whole WG
... but completely sympathetic to idea that all WG needs to be involved.

guus: let's leave it here and come back to it next time. hope to see some mailing list communication especially regarding potential use cases and requirements for SKOS

<RalphS> [I expect that we will want to create subcommittees in this WG but I'd like to suggest that important decisions come back to the WG telecon frequently as the subcommittees make progress]

guus: AOB? thanks for joining, we are adjourned.