Chatlog 2011-01-27

From RDFa Working Group Wiki
Jump to: navigation, search

See CommonScribe Control Panel, original RRSAgent log and preview nicely formatted version.

14:54:27 <RRSAgent> RRSAgent has joined #rdfa
14:54:27 <RRSAgent> logging to
14:54:29 <trackbot> RRSAgent, make logs world
14:54:29 <Zakim> Zakim has joined #rdfa
14:54:31 <trackbot> Zakim, this will be 7332
14:54:31 <Zakim> ok, trackbot; I see SW_RDFa()10:00AM scheduled to start in 6 minutes
14:54:32 <trackbot> Meeting: RDFa Working Group Teleconference
14:54:32 <trackbot> Date: 27 January 2011
15:00:00 <Knud> Knud has joined #rdfa
15:00:11 <Zakim> SW_RDFa()10:00AM has now started
15:00:19 <Zakim> + +3539149aaaa
15:00:27 <Knud> zakim, I am aaaa
15:00:27 <Zakim> +Knud; got it
15:00:36 <manu1> Agenda:
15:00:38 <manu1> Chair: Manu
15:00:56 <manu1> Present: Ivan, Benjamin, Manu, Knud, MarkB, ShaneM, Nathan, Steven
15:00:59 <ivan> zakim, dial ivan-voip
15:00:59 <Zakim> ok, ivan; the call is being made
15:01:00 <Zakim> -Knud
15:01:00 <Zakim> +Knud
15:01:00 <Zakim> +Ivan
15:01:09 <markbirbeck> markbirbeck has joined #rdfa
15:01:30 <Zakim> +??P54
15:01:36 <manu1> zakim, I am ??P54
15:01:36 <Zakim> +manu1; got it
15:02:42 <manu1> zakim, who is on the call?
15:02:42 <Zakim> On the phone I see Knud, Ivan, manu1
15:03:29 <markbirbeck> zakim, code?
15:03:29 <Zakim> the conference code is 7332 (tel:+1.617.761.6200 tel:+ tel:+44.203.318.0479), markbirbeck
15:04:16 <ShaneM> ShaneM has joined #rdfa
15:04:19 <Zakim> + +200000aabb
15:04:24 <markbirbeck> zakim, i am aabb
15:04:24 <Zakim> +markbirbeck; got it
15:04:27 <Zakim> + +1.612.217.aacc
15:05:08 <manu1> zakim, mute knud
15:05:08 <Zakim> Knud should now be muted
15:05:39 <manu1> zakim, who is on the call?
15:05:39 <Zakim> On the phone I see Knud (muted), Ivan, manu1, markbirbeck, +1.612.217.aacc
15:06:02 <ivan> zakim, aacc is ShaneM 
15:06:02 <Zakim> +ShaneM; got it
15:06:22 <ivan> zakim, mute me
15:06:22 <Zakim> Ivan should now be muted
15:06:39 <ivan> scribenick: ivan
15:06:58 <ivan> manu1: is it necessary to discuss the issue of default profile
15:07:12 <ivan> ... this may be a good idea in discussing with html5
15:07:28 <ivan> ... let us do the editorial issues first
15:07:46 <ivan> manu1: shane, did you look at steven's editorial issues?
15:07:54 <manu1> Topic: Approving Editorial suggestions?
15:07:57 <manu1> 1) Approve editorial suggestions?
15:07:58 <ivan> ShaneM: yes I have
15:07:58 <manu1>    ISSUE-71: Shelley Power's LC comments
15:08:00 <manu1>
15:08:01 <manu1>
15:08:03 <manu1>    ISSUE-79: Integrate CURIE information
15:08:05 <manu1>
15:08:06 <Steven_> Steven_ has joined #rdfa
15:08:07 <manu1>    ISSUE-80: Integrate attribute information
15:08:09 <manu1>
15:08:11 <manu1>    ISSUE-81: Make declarative definition normative, procedural
15:08:11 <ivan> manu1: what do you think are they ok?
15:08:12 <manu1>              definition informative.
15:08:14 <manu1>
15:08:50 <Steven_> zakim, dial steven-617
15:08:50 <Zakim> ok, Steven_; the call is being made
15:08:51 <ivan> Topic: ISSUE-79: Editorial merging sections on CURIEs
15:08:52 <manu1>
15:08:53 <Zakim> +Steven
15:09:10 <ivan> ... take issue 79: to merge some curie information
15:09:13 <Steven_> Sorry for being late, I was on another call, and missed the time
15:09:14 <ivan> ... my reaction is no
15:09:42 <ivan> ... we need a free standing curie section which is not only rdfa
15:09:52 <ivan> ... merging the sections would be problematic
15:10:17 <manu1>
15:10:22 <ivan> ... section 3.8 is a historical section that mark wrote back in the day to justify curie-s
15:10:42 <ivan> ... it does not really say anything about them and it is not normative
15:10:48 <ivan> ... i would prefer to let them alone, too
15:11:03 <ivan> manu1: essentially, issue 79 suggestion is to leave that as it is
15:11:04 <Steven_> Iḿ OK with that
15:11:05 <ivan> ShaneM: yep
15:11:17 <ivan> (WG accepted)
15:11:28 <Zakim> -manu1
15:11:45 <Zakim> +[IPcaller]
15:12:02 <ivan> Topic: ISSUE-80: Editorial - Integrate attribute datatype information
15:12:11 <manu1> ISSUE-80 - Integrate attribute information -
15:12:11 <trackbot> ISSUE-80 Editorial - Integrate attribute information. Triage of Issue 75 - Part 2 notes added
15:12:23 <ivan> .... the commenter is concerned that normative definitions and datatypes are scattered all over the place
15:12:29 <ivan> ... it was not true, but it might be true now
15:12:46 <ivan> ... section 8 has a lot of info, but it does not define any datatype (section 5)
15:12:58 <ivan> ... my proposal is to make it so that section 5 is complete
15:13:10 <ivan> ... it defined the attributes and syntax
15:13:54 <ivan> ... section 7.4.4., which is part of a larger section on curie and uri processing, I would be happy to remove
15:14:03 <ivan> ... it is defined in section (or it should be)
15:14:15 <manu1> zakim, I am [IPcaller]
15:14:15 <Zakim> ok, manu1, I now associate you with [IPcaller]
15:14:23 <manu1> zakim, who is on the call?
15:14:23 <Zakim> On the phone I see Knud (muted), Ivan (muted), markbirbeck, ShaneM, Steven, [IPcaller]
15:14:24 <ivan> ... mark, is it o.k with you?
15:14:43 <ivan> markbirbeck: it is fine with me; it is a self contained part, though
15:15:03 <ivan> ... but it is indeed a summary
15:15:11 <ivan> ... it does not add anything
15:15:28 <ivan> ShaneM: as it stands now, you have to read it
15:15:39 <ivan> ... it does not say what that datatype says
15:15:44 <ivan> ... it is safe to remove it
15:15:58 <ivan> manu1: agree with that
15:16:06 <ivan> ... any objection to remove that section?
15:16:17 <ivan> markbirbeck: how do we feel about other sections?
15:16:27 <ivan> manu1: nobody complained about other sections...
15:16:36 <ivan> ... so, maybe we can look through those
15:16:45 <ivan> ... at present we do not have any issues about this
15:17:08 <ivan> ShaneM: mark, if you have a strong objection, my alternative is to fix 7.4.4
15:17:21 <ivan> ... right now it is a bit coloquial
15:17:39 <ivan> markbirbeck: I do not have a strong objection, but, eg, 7.4.2 does it look any better?
15:17:49 <ivan> ... jenni would like to have everything in one place
15:17:54 <ivan> ... which makes sense
15:18:07 <ivan> ... if it is possible to fix 7.4.4 rather, I would prefer this a bit
15:18:23 <ivan> ... I am happy either way, I let shane decide
15:18:40 <ivan> ShaneM: mark, I agree that 7.2.2 has the same problem as 7.2.4 has, it is imprecise
15:19:00 <ivan> ... fixing it would mean referencing the datatypes back to the absolute definitions
15:19:26 <ivan> ... we are not referencing it here, there is no tie
15:20:04 <ivan> manu: there is also something here that says to make 7.4.4. non normative and the other normative
15:20:14 <ivan> ... that approach goes into the next issue we are talking about
15:20:49 <ivan> ... shane, do you agree making these explanatory section non-normative? This ties in into the next section
15:20:58 <ivan> ShaneM: she wanted section 8 to be non-normative
15:21:14 <ivan> ... section 7 there is no section I would make non-normative, it is important for implementers
15:21:25 <ivan> .... section 8 is more something like a test suite
15:21:36 <ivan> ... it gave me a bunch of examples
15:21:56 <ivan> ... I would defer to mark on whether section 8 should be non-normative
15:22:16 <ivan> manu: I trust you, shane, to make the right decision 
15:22:52 <ivan> ... talking about issue 80, shane offers to point back to the datatypes from the prose
15:23:08 <ivan> ShaneM: the same for 7.4.4
15:23:14 <ivan> ... they both need those tie-back
15:23:21 <ivan> manu: any objection?
15:23:23 <ivan> ...
15:23:26 <ivan> (WG agreed)
15:23:27 <ivan> Topic: ISSUE-81: Make section 7 normative, section 8 non-normative
15:23:29 <manu1> ISSUE-81 Make declarative definition normative, procedural definition informative,
15:23:46 <ivan> Steven_: I agree with her about this
15:24:09 <ivan> ... if we agree that 7.5 and 8 are overlapping, I agree making one normative and the other informative, advise for implementations
15:24:20 <ivan> ShaneM: I said I would defer to mark...
15:24:34 <ivan> markbirbeck: ... but you hinted it is a good idea:-) 
15:24:40 <manu1>
15:24:55 <ivan> ... the original idea was that one was a friendly explanation of section 7
15:25:03 <ivan> ... if we are sure that everything is covered
15:25:23 <ivan> ... then shane's suggestion (section 8 is informative, section 7 normative) is fine
15:25:58 <ivan> manu: when I did my implementation than I just implemented the process
15:26:05 <ivan> ... and then looking at the examples
15:26:33 <ivan> ShaneM: we are making so many changes that we will have a 2nd last call:-)
15:26:47 <ivan> ... I am not worried about the change
15:27:03 <ivan> ... I will have to make a cleaner implementation before 2nd last call
15:27:14 <ivan> q+
15:28:06 <ivan> manu:  from a design standpoint this is the right thing to do, if we find an issue
15:28:08 <manu1> ack ivan
15:28:09 <ivan> ack ivan 
15:28:31 <manu1> Ivan: I used Section 7 almost exclusively for my implementation.
15:28:38 <manu1> Ivan: I used section 8 for checking my understanding.
15:28:57 <ivan> manu: any objection to follow shane's offer, section 8 non-normative?
15:28:58 <ivan> ....
15:29:02 <ivan> (WG accepted)
15:29:05 <ivan> zakim, mute me
15:29:05 <Zakim> Ivan should now be muted
15:29:15 <ivan> Topic: ISSUE-71: Last Call Comments from Shelley Powers
15:29:15 <ivan> manu1: last issue is Shelley's comments
15:29:20 <manu1>
15:29:24 <ivan> ... everything that she had was editorial 
15:29:32 <manu1> ISSUE-71
15:29:35 <ivan> ... and they were not as heavy as Jeni's 
15:29:46 <ivan> ISSUE-71?
15:29:46 <trackbot> ISSUE-71 -- RDFa Core 1.1 LC comments from Shelley Powers -- open
15:29:46 <trackbot>
15:30:02 <ivan> ShaneM: sorry, I did not have time to look at those, let us skip those
15:30:14 <ivan> Topic: ISSUE-78: Prefixes and terms in one default RDFa Profile for all Host languages
15:30:14 <ivan> ISSUE-78?
15:30:14 <trackbot> ISSUE-78 -- Should we have default prefixes and terms for host languages -- open
15:30:14 <trackbot>
15:30:29 <ShaneM>
15:31:10 <manu1> ISSUE-73?
15:31:10 <trackbot> ISSUE-73 -- The RDFa WG needs to determine how each RDFa Profile document is managed -- open
15:31:10 <trackbot>
15:31:15 <ivan> manu1: it is the whole issue of the default profile, what goes there, etc
15:31:20 <ivan> zakim, unmute me
15:31:20 <Zakim> Ivan should no longer be muted
15:31:31 <manu1> Ivan: There are several sub-issues here
15:31:45 <manu1> Ivan: Do we want a default profile in the first place?
15:32:01 <manu1> Ivan: Is the content of the default profile frozen at the time of REC?
15:32:21 <manu1> Ivan: Is there a community-driven mechanism that allows people to add to the default profile over time?
15:32:32 <manu1> Ivan: In case we have a community-driven mechanism, what is it?
15:32:38 <Steven_> q+
15:32:42 <manu1> Ivan: These are all related
15:34:20 <manu1> Ivan: Do we want to have Dublin Core, FOAF, prefixes defined in the default profile for RDFa?
15:34:55 <manu1> Manu: Do we want to have a single RDFa default profile for all languages?
15:35:35 <manu1> Ivan: Do prefixes defined in the default profile scale? What about UAs that can't cache the profiles?
15:35:56 <manu1> ack Steven_
15:37:30 <manu1> Steven: I think RDFa does the right thing - we allow caching... registries are problematic.
15:37:40 <manu1> Ivan: Authors sometimes don't put in the namespace declarations.
15:37:50 <manu1> I have a proposal:
15:38:02 <manu1> We have 1 RDFa default profile for all languages.
15:38:19 <manu1> We allow new prefixes to be registered up until RDFa Core 1.1 goes to REC.
15:38:23 <manu1> same with terms.
15:38:41 <manu1> We don't allow new prefixes to be added to the RDFa default profile document after REC.
15:38:57 <manu1> but we do allow items to be suggested for the next revision of RDFa Core.
15:39:31 <manu1> q+
15:39:42 <manu1> zakim, [IPcaller] is me
15:39:42 <Zakim> +manu1; got it
15:39:45 <manu1> q?
15:40:46 <manu1> For a vocabulary to be included in the default profile, it must exhibit
15:40:48 <manu1> at least the following:
15:40:49 <manu1> 1. Be long-lived, use a URL redirecting service, or be controlled by an
15:40:51 <manu1>    organization that could ensure that the vocabulary stay reachable
15:40:52 <manu1>    for 10+ years or more.
15:40:54 <manu1> 2. Be of general use to web developers (so, rdf, rdfa, xsd, foaf, dc
15:40:55 <manu1>    would make the cut... unsure about skos and owl).
15:40:57 <manu1> 3. Be well documented, designed well and in use by a community that
15:40:58 <manu1>    can demonstrate that the vocabulary will be maintained for 10+ years.
15:41:19 <ShaneM> q+ to discuss profile evolution
15:41:42 <ivan> manu1: these are all issues, high level thoughts from everybody?
15:42:02 <ivan> Manu: I have put in irc my own approach
15:42:11 <ivan> ... we used to talk about xml, svg, etc profiles
15:42:30 <ivan> ... but what ivan put in on the mailing list to have only one default profile
15:42:42 <ivan> ... that would simplify things, only one profile is relevant
15:42:50 <ivan> ... I think that is a good idea
15:43:10 <ivan> ... as far as community managed registry: I think it would be a massive headache to have something that works for everyone
15:43:30 <ivan> ... if we agree to have that, we have to talk to the players
15:43:42 <ivan> ... this should be fixed
15:43:56 <ivan> ... when the rec are published
15:43:59 <ivan> q+
15:44:11 <ivan> ... and have some sort of a mechanism to update?
15:44:31 <manu1> ack [IPcaller]
15:44:35 <markbirbeck> q+
15:45:42 <ivan> manu1: proposal would be to take one registry, update it every X years, but not absolutely dynamic
15:45:43 <manu1> ack shaneM
15:45:43 <Zakim> ShaneM, you wanted to discuss profile evolution
15:46:04 <ivan> ShaneM: you suggest that host languages would not have a default profiles
15:46:24 <ivan> manu1: we would have one default profile for all our languages
15:46:37 <ivan> ShaneM: that would not solve things
15:46:46 <ivan> ... we have no announcement mechanism
15:46:59 <ivan> ... I would modify your proposal to say that host languages cannot define their own profile
15:47:02 <manu1> ack ivan
15:47:40 <manu1> ack markbirbeck
15:48:14 <ivan> markbirbeck: one problem is to have a uri to profile that keeps changing
15:48:26 <ivan> ... one step would be to freeze a profile but also freeze the uri
15:48:27 <manu1>
15:48:31 <manu1>
15:48:40 <ivan> ... based on a date
15:48:44 <ivan> ... which could then be changed
15:48:49 <manu1>
15:48:57 <ivan> ... what people want is that the profile attribute would not be specified
15:49:18 <ivan> ... you then allow people to refer to a profile specificly
15:49:22 <ivan> q+
15:49:38 <manu1> I like that suggestion, Mark
15:49:46 <ivan> ... but we have the possibility to have  a default profile for a language
15:49:59 <ivan> ... or default value for the profile attribute is XXX
15:50:21 <ivan> ... that gets round the moving thing
15:50:31 <ivan> ... but we talk about caching
15:50:51 <manu1> I agree that we need to hardcode profiles into processors...
15:50:56 <ivan> ... but always thoughts is that much more likely scenario is that people will hard code profiles 
15:51:07 <ivan> ... and then if you take that in context
15:51:12 <manu1> you don't need to, but I think that's what most implementations are going to do.
15:51:27 <ivan> ... in the original version we had the idea of profiles referring to other profiles
15:51:44 <ivan> ... that would give a much more dynamic features
15:52:02 <ivan> ... drupal could create a profile that would aggregate another profile
15:52:04 <ivan> ... etc
15:52:09 <ivan> q+
15:52:26 <manu1> I have issues w/ recursively sucking in profiles.
15:52:28 <manu1> ack ivan
15:53:21 <manu1> Ivan: What this means is that every 2 years, W3C would open a new group to revise profiles.
15:53:28 <manu1> Ivan: I don't think that's feasible.
15:53:53 <markbirbeck> @manu: I would say that's because you're seeing profiles� as something to be loaded on the fly. :)
15:54:44 <manu1> Ivan: If we can issue a new profile every 2 years, we make noise about it, implementers will have to update their implementations (hardcoded or not)
15:55:04 <manu1> Ivan: I don't think we should bind these two things together - profiles having a dated URI vs. non-dated URI is a good idea.
15:55:16 <manu1> Ivan: That's slightly orthogonal, though.
15:55:26 <manu1> q?
15:55:44 <ivan> markbirbeck: I think that we should not use the latest URI, we should always require an explicit URI
15:55:53 <ivan> ... you do not gain the latest 
15:56:26 <manu1> zakim, who is on the phone?
15:56:26 <Zakim> On the phone I see Knud (muted), Ivan, markbirbeck, ShaneM, Steven, manu1
15:56:45 <ivan> ShaneM: mark, to expand on someting you said
15:56:56 <ivan> ... assuming we had dated URI-s as profiles as they evolve
15:57:21 <ivan> ... would you expect implementation to hardcode all of those, so that pages that have explicit references to one would get it
15:57:45 <ivan> markbirbeck: well... I am picturing that there is a new release of a new ontology
15:57:57 <ivan> ... then w3c creates a new profile with the new ontology
15:58:08 <ivan> ... the only reason an implementer might hard code it 
15:58:23 <ivan> ... it is only for if somebody uses that explicitly
15:58:37 <ivan> ... whether this finds its way into an rdfa language
15:58:45 <ivan> ... so we would just refer to the latest
15:58:57 <ivan> ... in the rdfa document
15:59:05 <ivan> ... implementers may choose how they do it
15:59:37 <ivan> manu1: we should have a super session of LC
15:59:49 <ivan> .. our list is growing and we shall fill up the whole of february
15:59:53 <manu1>
16:00:07 <ivan> ... please put up your availability
16:00:13 <ivan> ... hopefully close a lot of them
16:01:23 <manu1> Should we have one default profile for all RDFa languages? Any objections?
16:03:09 <manu1> Shane: No objection, do you think it will work for HTML WG?
16:04:46 <manu1> PROPOSAL: RDFa 1.1 will have one default profile for all Host languages.
16:05:10 <ShaneM> +1 - NOTE that it might be a 'default default profile'
16:05:15 <manu1> +1
16:05:24 <Knud> +1
16:05:24 <ivan> Ivan: +1
16:05:25 <Steven_> +0
16:05:31 <markbirbeck> +1
16:05:39 <manu1> RESOLVED: RDFa 1.1 will have one default profile for all Host languages.
16:08:40 <manu1> PROPOSAL: RDFa WG will define a set of prefixes and terms for the default profile, those prefixes/terms will be frozen at REC (a mechanism will be setup to update the default profile before RDFa Core goes to REC)
16:09:33 <ivan> PROPOSAL: RDFa WG will define a set of prefixes and terms for the default profile; a mechanism will be setup to update the default profile before RDFa Core goes to REC
16:10:26 <ivan> PROPOSAL: RDFa WG will define a set of prefixes and terms for the default profile; a mechanism will be proposed to update the default profile
16:10:41 <markbirbeck> +1
16:10:46 <ivan> Ivan: +1
16:10:49 <manu1> +1
16:11:50 <manu1> zakim, who is on the call?
16:11:50 <Zakim> On the phone I see Knud (muted), Ivan, markbirbeck, ShaneM, Steven, manu1
16:11:51 <Knud> +1
16:11:57 <ShaneM> +1.....
16:13:02 <manu1> RESOLVED: RDFa WG will define a set of prefixes and terms for the default profile; a mechanism will be proposed to update the default profile
16:14:18 <markbirbeck> What about: PROPOSAL: RDFa WG will define a profile with a set of prefixes and terms, and this profile will be referenced as the default profile. A mechanism will also be proposed for creating additional profiles which can be referenced as the default profile for future versions of RDFa.
16:16:12 <Steven_> +1
16:18:01 <Zakim> -Knud
16:18:06 <Knud> Knud has left #rdfa
16:18:54 <Zakim> -markbirbeck
16:19:15 <markbirbeck> @Ivan: What you seek is simply not possible.
16:19:46 <markbirbeck> And it would be a major mistake to try to achieve it by having a profile that can change at any time.
16:19:55 <webr3> just fyi, js3 has over 100 prefixes defined by default, and in the past two weeks I've had 4 requests to add more to the default profile, including one two seconds ago to add bibo
16:20:05 <manu1>
16:20:20 <markbirbeck> It makes caching "meaningless", since you can't reliably cache.
16:20:40 <manu1>
16:20:59 <markbirbeck> It makes hard-coding into phones and small devices meaningless, too.
16:21:00 <manu1> RDFa 1.1 => default profile =>
16:21:22 <markbirbeck> (BTW, not sure why I got dropped...the phone went dead on me.)
16:21:28 <manu1> RDFa 2.0 => default profile =>
16:21:39 <ShaneM> q+ to discuss process
16:21:46 <manu1> ack shanem
16:21:46 <Zakim> ShaneM, you wanted to discuss process
16:24:09 <manu1> q+
16:25:13 <webr3> RDFa 1.1 processors must recognise the following prefixes: x,y,z - this list is also available as an RDFa Profile here: http://..... (then repeat for each new spec)
16:26:34 <manu1> ack
16:26:39 <manu1> ack manu1
16:27:10 <manu1> we seem to be coming to some sort of consensus:
16:29:27 <manu1> We bind RDFa 1.1 profile to a dated URL.
16:30:59 <manu1> For example:
16:31:12 <manu1> That profile could be updated every 2+ years
16:31:27 <manu1> RDFa 1.1 is bound to that URL as the default profile
16:32:40 <manu1>
16:34:23 <Zakim> -manu1
16:34:24 <ivan> zakim, drop me
16:34:24 <Zakim> -ShaneM
16:34:24 <Zakim> Ivan is being disconnected
16:34:28 <Zakim> -Ivan
16:34:32 <Zakim> -Steven
16:34:34 <Zakim> SW_RDFa()10:00AM has ended
16:34:38 <Zakim> Attendees were +3539149aaaa, Knud, Ivan, manu1, +200000aabb, markbirbeck, +1.612.217.aacc, ShaneM, Steven