IRC log of swbp on 2004-09-17
Timestamps are in UTC.
- 11:36:04 [RRSAgent]
- RRSAgent has joined #swbp
- 11:36:59 [danbri-scribe]
- context: http://lists.w3.org/Archives/Public/public-swbp-wg/2004Sep/0005.html
- 11:37:04 [danbri-scribe]
- present: tom, alistair, libby, danbri
- 11:37:23 [danbri-scribe]
- tom: wanted to brainstorm re scoping draft
- 11:37:29 [danbri-scribe]
- ...has had some discussion
- 11:37:41 [danbri-scribe]
- ...want to step back from that, ask about folks gut feeling re scope
- 11:38:00 [libby]
- http://lists.w3.org/Archives/Public/public-swbp-wg/2004Sep/0006.html
- 11:38:06 [danbri-scribe]
- ...focus on few key examples, per michael S's urging
- 11:38:19 [danbri-scribe]
- ...i'd be delighted to offer Dublin Core as an eg of hopefully good practice on a number of points
- 11:38:29 [danbri-scribe]
- ...other key examples could help illustrate
- 11:38:38 [danbri-scribe]
- [i'd be happy offering foaf similarly]
- 11:38:45 [danbri-scribe]
- ...also want reality check re how to proceed
- 11:38:59 [danbri-scribe]
- ...start with a scoping draft?
- 11:39:24 [danbri-scribe]
- ...div up, assign ownership to 2-parag sections
- 11:39:36 [danbri-scribe]
- ...put together a rough draft w/ a few refs on that particular point
- 11:39:41 [danbri-scribe]
- ...as a way of roughing up a 1st draft note
- 11:39:51 [danbri-scribe]
- ...can't begin think about that before we have better sense of scope
- 11:39:59 [danbri-scribe]
- ...reactions, comments?
- 11:40:41 [danbri-scribe]
- danbri: FOAF too, which draws on the +/- of the DC experience
- 11:40:55 [danbri-scribe]
- alistair: would be good contrast DC style with a more lightweight approach
- 11:41:56 [danbri-scribe]
- danbri: also rdf-calendar, skos, foaf, wgs84 geo, ...
- 11:42:20 [danbri-scribe]
- tom: interesting to hear dc seen as non-lightweight, we're trying to be lightweight but in context of wanting long term longevity etc
- 11:42:31 [danbri-scribe]
- al: i don't know the dc management style in detail...
- 11:42:51 [danbri-scribe]
- ...my impression from reading about the relative management styles, but for something as important as DC, seems reasonable
- 11:43:06 [danbri-scribe]
- ...whereas foaf more openended, imaginative style, so diff management style makes sense
- 11:43:23 [danbri-scribe]
- tom: trying to be lighteight but not weightless!
- 11:43:33 [danbri-scribe]
- ...like idea of organising things around certain styles, purposes
- 11:43:38 [danbri-scribe]
- ...which is one creating a vocab
- 11:43:39 [danbri-scribe]
- ?
- 11:43:45 [danbri-scribe]
- ...to present a range from formal to v informal
- 11:45:27 [danbri-scribe]
- danbri: scale from big (wordnet etc) medium (dc etc) small (extensions to foaf)
- 11:45:35 [danbri-scribe]
- tom: i'd like to see work around 'namespace policies'
- 11:45:46 [danbri-scribe]
- ...eg. when we assign an identifier, its there forever (in DC)
- 11:45:56 [danbri-scribe]
- ...but we can tweak the semantics in an certain range
- 11:46:10 [danbri-scribe]
- ...but if we change it too much, impacting apps, it implies the assignment of a new uriref
- 11:46:28 [danbri-scribe]
- tom: it'd be useful to articulate the range of expectations are associated w/ the assignment of a uri or uriref
- 11:46:45 [danbri-scribe]
- ...a lot of those things are discussed in a recent paper from Andy Powell
- 11:46:46 [danbri-scribe]
- @@url?
- 11:47:01 [danbri-scribe]
- ...at a minimum, if this taskforce...
- 11:47:08 [danbri-scribe]
- ...talk of urirefs in a DC implementation context
- 11:47:15 [danbri-scribe]
- ...but if we could elab on those principles
- 11:47:25 [danbri-scribe]
- ...or differential acc to purpose, style
- 11:47:33 [danbri-scribe]
- [aside: http://rdfweb.org/topic/FOAFCommunityProcess]
- 11:47:48 [danbri-scribe]
- ...addressses the user who has a vocab and wants to assign refs to it
- 11:48:09 [danbri-scribe]
- ...examine issues it raises in light of other types of vocab
- 11:49:15 [danbri-scribe]
- danbri: is DC using RDFS/OWL domain/range stuff now?
- 11:49:20 [danbri-scribe]
- tom: see Andy's abstrct model
- 11:49:30 [danbri-scribe]
- ...things can be represented as strings, but also by uri
- 11:49:39 [danbri-scribe]
- ...uses abstraect model to contrast different encodings
- 11:50:45 [aliman]
- danbri-scribe: categories of change to a schema?
- 11:51:05 [aliman]
- ... e.g. definition changes vs. schema mecahnics changes
- 11:51:58 [aliman]
- q+
- 11:52:34 [danbri-scribe]
- tom: wouldn't want to jump too much into mechanics of schemas
- 11:52:52 [danbri-scribe]
- ...without first getting an overview of some of the really basic issues of what you're assigning an ID too
- 11:52:53 [danbri-scribe]
- to
- 11:53:01 [danbri-scribe]
- ...what kinds of reasonable expectations
- 11:54:19 [danbri-scribe]
- thx
- 11:54:37 [danbri-scribe]
- al: agreeing w/ danbri re detail, but stability schema used in FOAF and SKOS is a valuable thing...
- 11:54:47 [danbri-scribe]
- ...would like to see it more formalised/stable
- 11:54:52 [Tbaker]
- http://dublincore.org/2004/06/14/dcq
- 11:54:54 [danbri-scribe]
- ...as valuable tool for developing schemas
- 11:54:59 [Ralph]
- Ralph has left #swbp
- 11:55:04 [danbri-scribe]
- ...also interest in different types of changes that occur
- 11:55:06 [aliman]
- http://www.w3.org/2004/02/skos/change.rdf
- 11:55:13 [danbri-scribe]
- ...as is an issue w/ skos concept schemes
- 11:55:51 [danbri-scribe]
- tom: ok going deeper in this direction so long as we cover basic assumptions sufficiently
- 11:56:22 [danbri-scribe]
- ...say v clearly, 'in order to be useful in the sw, a vocab isn't just a set of urirefs that are declared, but should/must/etc be declared in a schema...
- 11:56:25 [danbri-scribe]
- ...which can be used
- 11:56:29 [danbri-scribe]
- ...thats an underlying assumption
- 11:56:33 [danbri-scribe]
- ...which we slip into
- 11:56:46 [danbri-scribe]
- q+ re business models etc
- 11:56:51 [danbri-scribe]
- q+ to witter re business models etc
- 11:56:55 [aliman]
- q-
- 11:57:03 [danbri-scribe]
- ...terms + policies underlying assignments
- 11:57:39 [danbri-scribe]
- ack danbri
- 11:57:39 [Zakim]
- danbri-scribe, you wanted to witter re business models etc
- 11:59:03 [danbri-scribe]
- danbri: <witter/>
- 11:59:22 [danbri-scribe]
- al: i'd phrase these things as best practices, as tom says, but talking about the +/- rather than should/must
- 12:00:34 [aliman]
- danbri-scribe: foaf ns now content-negotiable
- 12:01:00 [aliman]
- ... also desire for accessing history (CVS style) of schema changes
- 12:01:08 [aliman]
- i.e. W3C style
- 12:01:10 [aliman]
- &DC
- 12:01:20 [danbri-scribe]
- danbri: would love to compare techniques for recording decision record + change log for datestamped namespace snapshots
- 12:01:32 [danbri-scribe]
- tom: yes v helpful, DC uses w3c-style datestamps...
- 12:01:38 [danbri-scribe]
- ...we've gone ahead and done this in DC
- 12:02:01 [danbri-scribe]
- ...I'd like a confirmation from a group like this, that its good practice, before formalizing
- 12:02:06 [libby]
- [see http://rdfweb.org/topic/FoafGalwayBreakoutSessions 2nd day stability discussions based around an update to the FOAFCommunityProcess document]
- 12:02:17 [danbri-scribe]
- ...as head of DC Usage Board, I assign the URI as covered by the NS policy to the term 'concept']
- 12:02:25 [danbri-scribe]
- ...and assign a uri to that particular version
- 12:02:43 [danbri-scribe]
- ...so we issue a new term-version with its own term-version uri, but has uri that is covered by the ns policy
- 12:03:02 [danbri-scribe]
- ...so in effect there are 2 identifiers pointing to the concept
- 12:03:21 [danbri-scribe]
- ...has been done like this for a while, have had +ve feedback, but would be good to get a group like this to examine this practice
- 12:03:49 [danbri-scribe]
- ...would be great if it did turn out to be good practice
- 12:04:03 [danbri-scribe]
- danbri: are Usage Board proceedings public now?
- 12:04:11 [danbri-scribe]
- tom: yes, you can get at it thru the meeting pages
- 12:04:29 [danbri-scribe]
- ...you can get at it thru the decision records, which are linked to the supporting materials, etc.
- 12:04:38 [danbri-scribe]
- ...decision text linked
- 12:04:45 [danbri-scribe]
- ...linked to an issuing of a new term version
- 12:05:08 [aliman]
- q+ to talk about term / scheme versioning
- 12:05:08 [danbri-scribe]
- ...in w3c datestamp style
- 12:05:14 [danbri-scribe]
- ...analgous to w3c approach, ...
- 12:05:29 [danbri-scribe]
- ...i've been looking for a while for a specific sort of spec of the w3c approach
- 12:05:50 [danbri-scribe]
- [@@ pubrules + TR-in-rdf urls]
- 12:06:14 [danbri-scribe]
- ...could slice up the attributes of the problem differently
- 12:07:02 [danbri-scribe]
- al: i think best practices for this very in scope for this TF
- 12:07:20 [danbri-scribe]
- ...from my pov its interesting to figure out if DC's term versioning approach is right thing for large thesauri
- 12:07:29 [danbri-scribe]
- ...which often have their own internal approaches
- 12:07:50 [danbri-scribe]
- +1 on versioning being in scope (esp for RDF namespaces)
- 12:07:56 [danbri-scribe]
- http://lists.w3.org/Archives/Public/public-swbp-wg/2004Sep/0005.html
- 12:08:02 [danbri-scribe]
- tom: can we walk thru the draft?
- 12:08:05 [libby]
- http://lists.w3.org/Archives/Public/public-swbp-wg/2004Sep/0006.html
- 12:08:12 [libby]
- is more detailed
- 12:08:50 [danbri-scribe]
- reviewing 0006...
- 12:08:57 [danbri-scribe]
- tom: wanted to keep to 15-20 pages
- 12:09:00 [danbri-scribe]
- ...at most if poss
- 12:09:18 [danbri-scribe]
- ...describe principles of BP
- 12:09:23 [danbri-scribe]
- ...describe a range of styles
- 12:09:28 [danbri-scribe]
- ...can have length bibliog
- 12:09:52 [danbri-scribe]
- ...would try to create a short readable document which introduces ppl to the range of good practice solns that we see
- 12:10:06 [danbri-scribe]
- Q2: about specific outline ...
- 12:10:21 [danbri-scribe]
- ...we prob need to define a few basic terms, like 'vocabulary', and 'term'
- 12:10:47 [danbri-scribe]
- i'd love to see a clear statement about why we say uriref not uri (@@rfc2396bis redraft will do this in appendix)
- 12:10:58 [danbri-scribe]
- ...why talk about namespaces, is it a tricky term, and why?
- 12:11:05 [danbri-scribe]
- ...section 2, need to characterise the scope
- 12:11:18 [danbri-scribe]
- ...are we talking about metadata elements sets, controlled vocabs, taxonomies
- 12:11:24 [danbri-scribe]
- q+ to suggest priorities
- 12:11:31 [aliman]
- q-
- 12:11:32 [danbri-scribe]
- ...and articulate some basic assumptions re SW
- 12:11:43 [danbri-scribe]
- ...if our goal is to present a range of styles for ppl to decare vocabs for sw
- 12:11:58 [danbri-scribe]
- ...are we expecting formal schema declarations
- 12:12:04 [danbri-scribe]
- ...articulate underlying assumptions
- 12:12:11 [danbri-scribe]
- ...shouldn't slip into a generic intro-to-sw
- 12:12:19 [danbri-scribe]
- ...some danger, but need articulate our assumptions
- 12:12:25 [danbri-scribe]
- ...section 3., principles of good practice
- 12:12:31 [danbri-scribe]
- ...has some strawman content there
- 12:12:44 [danbri-scribe]
- ...happy to go as far as we can to articulate/agree on indiv principles
- 12:13:02 [danbri-scribe]
- q+ to ask if UB meeting at DC would accept an observer
- 12:13:24 [danbri-scribe]
- ...am happy to wordsmith etc., get into coherent style, would divide up the work of assembling the sets of args related to particular pricniples
- 12:13:56 [danbri-scribe]
- ...for any issues that we recognise to be important, but too unsettled/immature, where no emergent best practice yet., push them into a Section 4
- 12:14:19 [danbri-scribe]
- ...would try characterise the debate on particular issues
- 12:14:29 [danbri-scribe]
- ...point ppl on work that's going on, eg. around metadata registries
- 12:14:37 [danbri-scribe]
- [(aka 'directories')]
- 12:14:50 [danbri-scribe]
- ...i'd not try to do too much there as out of scope for our central goals
- 12:15:08 [danbri-scribe]
- ...point ppl at registry initiatives, summarise issues but point off elsewhere
- 12:15:26 [danbri-scribe]
- pt 3; re section1 terminology
- 12:15:39 [danbri-scribe]
- we need some vague sense of what a term is
- 12:16:00 [danbri-scribe]
- ...'there is a notion of term' point to diff approaches
- 12:16:10 [danbri-scribe]
- ...i've tried to get consensus on this before, can be costly
- 12:18:29 [aliman]
- danbri-scribe: for ground up projects like FAOF RSS SKOS possible to talk about RDF worldview ...
- 12:18:47 [aliman]
- ... less easy for established systems migrating to RDF
- 12:19:13 [danbri-scribe]
- al: what if we have a glossary for this document, but appendix w/ defs we're using
- 12:19:24 [danbri-scribe]
- ...and focus of the doc is more on illustration by example
- 12:19:25 [danbri-scribe]
- +1
- 12:19:29 [danbri-scribe]
- tom: yes agree w/ that
- 12:19:31 [libby]
- +1
- 12:19:40 [danbri-scribe]
- ...also Andy's abstract model, he initially started w/ upfront definition of terms
- 12:19:46 [danbri-scribe]
- ...but I see he's moved to appendix
- 12:20:24 [danbri-scribe]
- ...we just need to make sure the terms are introduced well enough to make things interpretable
- 12:20:32 [danbri-scribe]
- ...i don't see the definitions are the major contrib of the paper
- 12:20:43 [danbri-scribe]
- ...just as a way of making sure we are clarifying what it is we're talking about
- 12:20:52 [danbri-scribe]
- ...when we say 'term' this is roughly what we mean
- 12:21:28 [danbri-scribe]
- ...things like ownership of namespace, or of a vocab., attracted discussion on BP list earlier
- 12:21:46 [danbri-scribe]
- ...if we can capture how that's controversial, would be a good contrib
- 12:21:54 [danbri-scribe]
- ...should we give things a handle, eg 'namespace owner'
- 12:21:56 [danbri-scribe]
- ?
- 12:22:03 [danbri-scribe]
- ...we shouldnt start by trying to define terms
- 12:22:13 [danbri-scribe]
- ...but make sure, when we finish, that we've defined them sufficiently
- 12:22:32 [danbri-scribe]
- [5 mins left]
- 12:22:48 [danbri-scribe]
- ...next section, re underlying assumptions
- 12:23:02 [danbri-scribe]
- ...open, loosly coupled, mixed-language environment
- 12:23:17 [danbri-scribe]
- (@@rdf concepts + primer, do they cover enough of it?)
- 12:23:34 [danbri-scribe]
- ...point 3 in scoping draft
- 12:23:42 [danbri-scribe]
- pt 4 re section 2
- 12:24:17 [aliman]
- q+ to suggest a different order
- 12:25:05 [danbri-scribe]
- q-
- 12:25:13 [danbri-scribe]
- ah, I think http://rdfweb.org/topic/FOAFCommunityProcess might fit into this section someho
- 12:25:14 [danbri-scribe]
- w
- 12:25:24 [danbri-scribe]
- [[
- 12:25:24 [danbri-scribe]
- 4. About Section 3 - Principles of Good Practice: This is the
- 12:25:24 [danbri-scribe]
- part which (I would hope) could form the core contribution
- 12:25:24 [danbri-scribe]
- of the VM note,
- 12:25:25 [danbri-scribe]
- ]]
- 12:25:39 [libby]
- q
- 12:25:42 [danbri-scribe]
- q?
- 12:25:52 [libby]
- q+
- 12:25:54 [danbri-scribe]
- tom, prob of resolving/derferencing uris
- 12:25:58 [danbri-scribe]
- ...who is recommending what?
- 12:26:01 [danbri-scribe]
- ...eg. DC practice
- 12:26:21 [danbri-scribe]
- ...would follow whatever evolving practice is
- 12:26:31 [danbri-scribe]
- ...deref an term URI you get RDF schema
- 12:26:35 [danbri-scribe]
- ...also relevant TAG finding
- 12:27:00 [danbri-scribe]
- ...assuming we unlikely to agree on a particular recommendation
- 12:27:16 [danbri-scribe]
- ....describe range of solutions
- 12:27:24 [aliman]
- Dereferencing bets practises will depend on size of vocab
- 12:27:33 [danbri-scribe]
- ...what are the attributes of a term?
- 12:27:49 [aliman]
- (But like dereferencing in scope)
- 12:27:53 [danbri-scribe]
- ...to what extent are there stds/models that one can/should follow?
- 12:28:32 [danbri-scribe]
- I'd propose RDFS/OWL as one strong model of term attributes
- 12:30:31 [aliman]
- danbri-scribe: would be nice to say: these are the things useful to include in term (i.e. class & prop) descriptions.
- 12:30:49 [libby]
- [in case we dont't come to it: advice oon stability is important to me; also validation, app profiles; also there are a number of v short one-line questions that have come up re owl]
- 12:30:53 [aliman]
- ... re. RDFS/OWL vocabs
- 12:30:58 [danbri-scribe]
- http://rdfweb.org/topic/FOAFCommunityProcess
- 12:32:02 [libby]
- [also how far we go into managing a community process we go into, e.g the rdf-calendar stuff for managing consensus]
- 12:33:01 [danbri-scribe]
- tom: i see this in a versioning context, having a timestamped and a non-timestamped, is that a translation can refer to a timestamped URI but be about a generic URI
- 12:33:05 [danbri-scribe]
- ...so you can have the link to both
- 12:33:26 [danbri-scribe]
- ...concept thats used in assertions, and the specific historical version
- 12:33:31 [danbri-scribe]
- ...hoping we can get that far
- 12:33:56 [danbri-scribe]
- q?
- 12:34:05 [danbri-scribe]
- ack aliman
- 12:34:05 [Zakim]
- aliman, you wanted to suggest a different order
- 12:34:29 [danbri-scribe]
- al, re 0006.html sections2/3/4, i like what you've got in Scope there
- 12:34:43 [danbri-scribe]
- ...actually i'd put section 2 last, or at least leave it., and try with this document to focus on 3 first
- 12:34:52 [danbri-scribe]
- ...bring that up top, to make it easier to get into
- 12:35:02 [libby]
- makes sense to me
- 12:35:07 [danbri-scribe]
- ...could turn out to be v widely read and important, so bring the substance up to top
- 12:35:12 [danbri-scribe]
- ack libby
- 12:35:18 [danbri-scribe]
- [general agreement]
- 12:35:29 [danbri-scribe]
- libby: you see web developers/hackers trying to get into ontology creation
- 12:35:38 [danbri-scribe]
- tom: adjourning...
- 12:35:58 [danbri-scribe]
- ...suggest over-discussion of scope would be counter productive
- 12:36:16 [danbri-scribe]
- ...start assigning sections, see how things evolve
- 12:36:37 [danbri-scribe]
- ...general agreement w/ rough scope
- 12:36:38 [danbri-scribe]
- +1
- 12:36:44 [aliman]
- +1
- 12:36:49 [libby]
- +1
- 12:36:57 [danbri-scribe]
- tom: happy moving terminology to back of doc
- 12:36:58 [libby]
- I like it a lot
- 12:37:24 [danbri-scribe]
- ...yesterdays swbp telecon suggested 1 more round of comments would be useful on scope draft
- 12:37:30 [danbri-scribe]
- ...but i could work with what's come in so far
- 12:37:43 [danbri-scribe]
- ...turn it back into an outline w/ sections reordered as discussed here
- 12:37:58 [danbri-scribe]
- ...i'd somehow want to make leap between that and assigning sections to ppl
- 12:38:12 [danbri-scribe]
- ...can try to come up w/ a draft
- 12:38:21 [danbri-scribe]
- ...outline is 3 or 4 pages already
- 12:41:21 [danbri-scribe]
- [discussion of wiki vs html vs xyz for collab]
- 12:42:11 [danbri-scribe]
- tom, see http://esw.w3.org/topic/FrontPage
- 12:42:59 [aliman]
- e.g. http://esw.w3.org/topic/SkosDev
- 12:44:13 [Zakim]
- -aliman
- 12:44:20 [danbri-scribe]
- ---------
- 12:56:12 [Tbaker]
- http://www.bi.fhg.de/People/Thomas.Baker/ISSUES/
- 13:24:10 [Zakim]
- -Tbaker
- 13:24:11 [Zakim]
- -bristol
- 13:24:12 [Zakim]
- SW_BPD(VM)7:30AM has ended
- 13:24:13 [Zakim]
- Attendees were Tbaker, aliman, libby, danbri
- 13:45:25 [TomAdams]
- TomAdams has joined #swbp
- 13:47:59 [TomAdams]
- TomAdams has left #swbp
- 14:00:33 [danbri_dna]
- danbri_dna has joined #swbp
- 14:01:10 [danbri_dna]
- danbri_dna has left #swbp
- 14:20:23 [TomAdams]
- TomAdams has joined #swbp
- 14:20:30 [TomAdams]
- TomAdams has left #swbp
- 14:45:01 [Zakim]
- Zakim has left #swbp
- 15:35:28 [TomAdams]
- TomAdams has joined #swbp
- 15:35:34 [TomAdams]
- TomAdams has left #swbp