Chatlog 2010-05-27

From RDFa Working Group Wiki
Jump to: navigation, search

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

13:33:14 <RRSAgent> RRSAgent has joined #rdfa
13:33:15 <RRSAgent> logging to
13:52:29 <manu> trackbot, prepare telecon
13:52:31 <trackbot> RRSAgent, make logs world
13:52:33 <trackbot> Zakim, this will be 7332
13:52:33 <Zakim> ok, trackbot; I see SW_RDFa()10:00AM scheduled to start in 8 minutes
13:52:34 <trackbot> Meeting: RDFa Working Group Teleconference
13:52:34 <trackbot> Date: 27 May 2010
13:53:13 <markbirbeck> markbirbeck has joined #rdfa
13:53:23 <manu> Agenda:
13:53:24 <ShaneM> ShaneM has joined #rdfa
13:53:28 <manu> Chair: Manu
13:53:41 <manu> Present: Manu, Shane, MarkB, Benjamin, Ivan, Shane, Toby, Knud, Steven
13:54:01 <manu> Regrets: BenA
13:55:23 <Steven> Steven has joined #rdfa
13:58:13 <tinkster> tinkster has joined #rdfa
13:59:13 <Zakim> SW_RDFa()10:00AM has now started
13:59:20 <Zakim> +tinkster
13:59:24 <tinkster> zakim, mute me
13:59:24 <Zakim> sorry, tinkster, muting is not permitted when only one person is present
13:59:33 <Zakim> +[IPcaller]
13:59:36 <tinkster> zakim, why's that?
13:59:36 <Zakim> I don't understand your question, tinkster.
13:59:40 <manu> zakim, +IPcaller is manu
13:59:40 <Zakim> sorry, manu, I do not recognize a party named '+IPcaller'
13:59:42 <tinkster> zakim, mute me
13:59:42 <Zakim> tinkster should now be muted
13:59:43 <ivan> zakim, dial ivan-vip
13:59:43 <Zakim> I am sorry, ivan; I do not know a number for ivan-vip
13:59:45 <manu> zakim, IPcaller is manu
13:59:45 <Zakim> +manu; got it
13:59:54 <ivan> zakim, dial ivan-voip
13:59:54 <Zakim> ok, ivan; the call is being made
13:59:55 <Zakim> +Ivan
13:59:59 <tinkster> zakim, unmute me
14:00:00 <Zakim> tinkster should no longer be muted
14:00:50 <Steven> zakim, dial steven-617
14:00:50 <Zakim> ok, Steven; the call is being made
14:00:52 <Zakim> +Steven
14:00:58 <tinkster> mhausenblas just posted this link on #swig -
14:01:19 <Zakim> +McCarron
14:01:34 <Zakim> +[MIT528]
14:02:10 <Benjamin> Benjamin has joined #rdfa
14:02:11 <Knud> Knud has joined #rdfa
14:02:27 <Zakim> +Knud
14:02:29 <dongmei> dongmei has joined #rdfa
14:03:20 <markbirbeck> zakim, code?
14:03:20 <Zakim> the conference code is 7332 (tel:+1.617.761.6200 tel:+ tel:+44.117.370.6152), markbirbeck
14:03:22 <Zakim> +BenjaminA
14:04:09 <Steven> zakim, who is noisy?
14:04:20 <Zakim> Steven, listening for 10 seconds I heard sound from the following: manu (44%), Ivan (44%), Knud (24%), BenjaminA (4%)
14:04:49 <ivan> zakim, mute Knud 
14:04:49 <Zakim> Knud should now be muted
14:05:04 <Zakim> +markbirbeck
14:05:29 <Steven> scribe: Steven
14:05:45 <Steven> rrsagent, make minutes
14:05:45 <RRSAgent> I have made the request to generate Steven
14:06:04 <Steven> rrsagent, make log public
14:06:36 <Steven> zakim, who is here?
14:06:36 <Zakim> On the phone I see tinkster, manu, Ivan, Steven, McCarron, [MIT528] (muted), Knud (muted), BenjaminA, markbirbeck
14:06:38 <Zakim> On IRC I see dongmei, Knud, Benjamin, tinkster, Steven, ShaneM, markbirbeck, RRSAgent, Zakim, manu, kennyluck, ivan, trackbot
14:07:06 <Steven> Topic: website
14:07:38 <Steven> Tinkster: [describes rdfa checker]
14:08:02 <ivan> -> entry about check.rdfa on the tool list at W3C
14:08:10 <tinkster>
14:08:12 <Steven>
14:08:32 <Steven> tinkster: Not handling non-ascii well yet
14:08:40 <manu>
14:09:15 <Steven> Manu: I ran it on newsweek; cool!
14:09:44 <Steven> tinkster: The display is valid rdfa too
14:10:15 <Steven> Mark: Hmm, run the output through itself?
14:10:17 <Knud>
14:10:31 <Steven> Steven: Lol
14:10:33 <Knud> Internal Server Error... :)
14:11:20 <Steven> Manu: Live editing mechanism?
14:11:27 <markbirbeck> Does it handle redirects? Often useful to pass shortened URLs in.
14:11:30 <Steven> ... to generate rdfa for the user
14:12:10 <Steven> ... quick and dirty way to generate rdfa
14:12:24 <Steven> Steven: Yes it handles redirects
14:13:05 <Steven> Manu: anyone working on that?
14:13:17 <tinkster> markbirbeck, it should handle redirects i think.
14:13:26 <Steven> Steven: Put on as a posting?
14:13:37 <Steven> Manu: I don't have access
14:13:45 <Steven> Mark: Huh? You should!
14:13:58 <Steven> ... Toby, do you?
14:14:05 <Steven> Toby: No.
14:14:41 <Steven> Mark: I shall ping Ben, to add everyone
14:15:16 <Steven> ACTION: Mark to add people to
14:15:16 <trackbot> Created ACTION-30 - Add people to [on Mark Birbeck - due 2010-06-03].
14:15:29 <Steven> Topic: RDFa DOM API progress
14:15:31 <manu>
14:15:44 <Steven> Manu: I have made changes based on Ivan's feedback
14:16:10 <Steven> ... now we have intro, basic concepts, and that is enough to start using it
14:16:17 <Steven> ... the rest is detail
14:16:36 <Steven> ... I did skim over Ivan's email. Ivan is this good for you?
14:17:02 <Steven> Ivan: I only have small problems. I sent another mail an hour or so ago
14:17:07 <Steven> ... please see that
14:17:12 <Steven> ... no major issues
14:17:35 <Steven> ... biggest issue is that I got confused between "high" and "low"
14:18:12 <Steven> Manu: Any other comments?
14:18:51 <Steven> ... FPWD in a week?
14:19:07 <Steven> ... I'll send to the group Monday, and a strawpoll next Thursday
14:19:37 <Steven> ... By the way, there are new concept diagrams
14:19:50 <Steven> ... and a couple of added sections
14:19:57 <Benjamin> q+
14:20:09 <manu> ack Benjamin
14:20:43 <Steven> Benjamin: It reads fluently. I agree with Ivan's restructuring ideas, minor issues; really cool. See my mail
14:21:25 <Steven> Topic: Processing Behavior when @profile document is not available
14:21:33 <Steven> issue-25?
14:21:33 <trackbot> ISSUE-25 -- Halt generation of triples in sub-trees when a @profile document cannot be retrieved -- open
14:21:33 <trackbot>
14:22:00 <manu>
14:22:31 <Steven> Toby: We allow profiles that define prefixes and keywords.
14:22:47 <Steven> .... If the profile is not retrievable or other problems
14:23:03 <Steven> ... then we can't be sure how to interpret children
14:23:49 <Steven> ... in that case we should ignore child elements
14:24:05 <Steven> ... Ivan and I disagree about what to do with incomplete triples
14:24:19 <Steven> ... I say use bnodes, he says don't complete
14:24:27 <manu> q+
14:24:32 <Steven> ... I don't have a strong opinion
14:24:38 <Steven> Ivan: Simplicity!
14:25:20 <manu> q-
14:25:25 <Steven> Toby: We can complete with a relative URI, but not absolute, because we don't know if it is a CURIE or not
14:25:29 <manu> Manu: Yes, I agree - simplicity!
14:25:30 <markbirbeck> Q+
14:25:42 <manu> ack markbirbeck
14:25:53 <Steven> Mark: Why if it's relative can we distinguish?
14:25:59 <Steven> ... it might be a token
14:26:08 <Steven> Toby: If it's got a slash or something....
14:26:28 <Steven> Mark: This helps only for processing
14:26:51 <ivan> q+
14:26:57 <Steven> ... why not create a new named graph for the stuff you're not sure about
14:27:10 <Steven> ... then you can come back later and fix up
14:28:30 <Steven> Steven: Can't users who need to do that just do it?
14:28:56 <Steven> Mark: I'm worried about text preventing people from doing it
14:29:14 <Steven> ... if it says you must fail, then you can't fix up
14:29:26 <manu> ack ivan
14:29:30 <Steven> Toby: I'm OK with wording that allows both
14:29:55 <Steven> Ivan: Named graphs are fine, but in practice most processors produce one graph
14:30:05 <Steven> ... no standard for named graphs
14:30:19 <markbirbeck> q+
14:30:45 <manu> ack markbirbeck
14:30:50 <Steven> ... so it shouldn't be required to produce a named graph
14:31:26 <Steven> Mark: I didn't mean to enforce it. Just allow it as an option for processing
14:31:37 <ShaneM> The spec currently says this:
14:31:37 <ShaneM>       This specification does not say anything about what should happen to the triples generated, or whether more
14:31:37 <ShaneM>       triples might be generated during processing than are outlined here. However, to be conforming, an
14:31:37 <ShaneM>       RDFa processor MUST act as if at a minimum the rules in this section are applied, and a single
14:31:37 <ShaneM>       <tref>RDF graph</tref> produced. As described in the <a href="#processorconf">RDFa Processor Conformance</a> section,
14:31:38 <Steven> ... I have tons of named graphs in my processor
14:31:40 <ShaneM>       any additional triples generated MUST NOT appear in the <tref>default graph</tref>.
14:32:21 <Steven> Mark: People should be allowed to do this
14:32:26 <tinkster> q+
14:32:31 <Steven> Ivan: I don't have a problem with that
14:32:33 <manu> ack tinkster
14:32:56 <Steven> Toby: You can't always tell what is missing from a given profile
14:33:06 <Steven> Mark: Yes
14:33:40 <Steven> Ivan: If the profile doesn't work, nothing more is generated in the default graph seems like a good position
14:33:44 <Steven> Mark: Yeah
14:34:13 <Steven> Shane: I'm happy to put that in, but even if we don't, it is still allowed (see the quote above)
14:35:00 <markbirbeck> Not quite...current proposal is this:
14:35:05 <markbirbeck> "The suggested solution, originally from Jeni Tennison is that when
14:35:05 <markbirbeck> parsers encounter a profile URI that cannot be retrieved, they should
14:35:05 <markbirbeck> not process that subtree."
14:35:05 <Steven> Manu: So all incomplete triples are not added to the default graph?
14:35:42 <Steven> Toby: Sibling elements are OK, so don't clear incomplete triples
14:36:29 <tinkster> q+ to ask about non-RDFa, but XHTML profiles.
14:37:31 <manu> PROPOSAL: When a @profile cannot be retrieved, all incomplete triples cannot be completed by the element that contains the non-retrievable @profile. The subtree containing the non-retrievable @profile must not be processed.
14:37:34 <Steven> Toby: Just treat it as if the element wasn't there
14:37:46 <tinkster> +1
14:37:58 <Steven> Steven: should not?
14:38:42 <Steven> Mark: No more triples should be added to the default graph
14:38:49 <Steven> ... for that subtree
14:39:07 <manu> PROPOSAL: When a @profile cannot be retrieved, all incomplete triples that cannot be completed by the element that contains the non-retrievable @profile. The subtree containing the non-retrievable @profile must not generate triples to be placed into the default graph.
14:39:19 <tinkster> +1
14:39:42 <tinkster> (the first sentance doesn't make much sense)
14:39:55 <manu> PROPOSAL: When a @profile cannot be retrieved, all incomplete triples must not be placed in the default graph. The subtree containing the non-retrievable @profile must not generate triples to be placed into the default graph.
14:39:59 <ShaneM> ... the current element and its children must not cause triples to be placed into the default graph.
14:40:52 <manu> PROPOSAL: When a @profile cannot be retrieved, the current element and its children must not cause triples to be placed into the default graph.
14:41:39 <manu> ack tinkster
14:41:39 <Zakim> tinkster, you wanted to ask about non-RDFa, but XHTML profiles.
14:41:54 <ShaneM> q+ to discuss profile parsing
14:42:14 <Steven> Toby: If we retrieve the profile and it has zero triples, is in another format
14:42:18 <Steven> Ivan: Should be OK
14:42:23 <Steven> Toby: Edge case?
14:42:29 <Steven> Ivan: No.
14:42:46 <Steven> Ivan: Profiles can be used for other things too
14:43:51 <manu> ack shanem
14:43:51 <Zakim> ShaneM, you wanted to discuss profile parsing
14:43:51 <Steven> ... the error case is only for non-retrievable, and or non-syntactic profiles
14:44:02 <ivan> q+
14:44:25 <Steven> Shane: I don't think we can legislate this
14:44:38 <Steven> ... the term is "available"
14:44:41 <manu> PROPOSAL: When a @profile is not available, the current element and its children must not cause triples to be placed into the default graph.
14:44:50 <Steven> ... if the profile is not available, then problem
14:45:06 <markbirbeck> q+
14:45:07 <Steven> ... if it is available, and produces no terms, then that is not a problem
14:45:11 <Steven> q+
14:45:52 <Steven> Ivan: If I get a profile with a bug, that is a bug, so we can't legislate
14:46:12 <ivan> ack ivan
14:46:39 <Steven> Mark: I don't think we can resolve this on the phone, but we need somnething like "when profile is not available"
14:46:41 <manu> ack markbirbeck
14:46:54 <ShaneM> PROPOSED:            If any referenced RDFa Profile is not available, then the <tref>current element</tref> and its children must not           place any triples in the <tref>default graph</tref>.
14:47:08 <Steven> ... Javascript parser can't go to different domains, so it might have a hardwired list
14:47:22 <Steven> ... and 'available' might mean something else
14:47:25 <manu> ack steven
14:47:34 <tinkster> If we require that all RDFa profiles contain this triple: { <> rdf:type rdfa:Profile . }, then once a profile is retrieved, we can verify that it has worked OK.
14:47:42 <ivan> +1 to steven
14:47:47 <ShaneM> +1 to steven
14:47:52 <manu> +1
14:48:09 <ShaneM> +1 to tinkster too - that's clever
14:48:18 <Steven> Steven: A profile that has no additions is like zero. No reason to disallow it
14:48:42 <ivan> +1
14:48:45 <tinkster> +1
14:48:48 <Benjamin> +1
14:48:57 <manu> +1
14:49:00 <Steven> +1
14:49:09 <Knud> +1
14:49:23 <markbirbeck> +1
14:49:25 <Steven> [+1s are to Shanes proposal above]
14:49:33 <ivan> q+
14:49:34 <tinkster> I'd like to come up with some examples of available-but-unusable profiles; I'll try to post them to the mailing list soon.
14:49:37 <ivan> (side issue)
14:50:02 <manu> RESOLVED: If any referenced RDFa Profile is not available, then the "current element" and its children must not place any triples in the "default graph".
14:50:58 <Steven> Ivan: I will answer Jenni officially
14:51:13 <Steven> ACTION: Ivan to reply to Jenny about @profile unavailable
14:51:13 <trackbot> Created ACTION-31 - Reply to Jenni about @profile unavailable [on Ivan Herman - due 2010-06-03].
14:51:43 <manu> Topic: Profile attribute value processing order
14:51:45 <manu>
14:52:04 <tinkster>
14:52:09 <Steven> issue-23?
14:52:09 <trackbot> ISSUE-23 -- @profile order - RDFa Core 1.1 versus HTML 4.01 and XMDP -- open
14:52:09 <trackbot>
14:52:27 <Steven> Toby: HTML4 says only the first URI in the list is used
14:52:29 <manu> Toby's e-mail proposal for ISSUE-23:
14:53:24 <Steven> Toby: Solution to process the URIs in reverse order
14:53:28 <ivan> +1
14:53:29 <ShaneM> +1 to this solution
14:53:40 <Steven> Steven: Fine
14:53:55 <Steven> Mark: You don't have to describe it as processing in reverse order
14:54:47 <Steven> Manu: It doesn't necessarily match with CSS and Javascript evaluation
14:55:02 <Steven> Mark: I think we're covered, see my mail
14:55:41 <Steven> ... the L to R/R to L is bound to the language itself
14:56:08 <Steven> ... programmers aren't expecting a general rule
14:56:30 <Steven> Manu: When shall we resolve this?
14:56:33 <Steven> Ivan: Now
14:56:54 <Steven> Mark: Let Manu roll it over, since he's the main objector
14:57:08 <Steven> Manu: I'd be happier to leave it
15:00:02 <Steven> Steven: I am talking on RDFa at next week, ~600 attendees
15:00:16 <Steven> Ivan: I am in Seattle soon, giving a talk on RDFa
15:00:36 <Steven> Shane: Whitespace resolution from last week
15:00:37 <ShaneM>
15:01:15 <Steven> Ivan: Core defines what whitespace is for RDFa
15:02:01 <Steven> Shane: HTML5 has a intricate definition for whitespace
15:02:33 <Steven> Shane: We should watch for that
15:03:01 <Steven> Manu: Not a danger; HTML5 is not the only language that uses RDFa
15:03:20 <ShaneM>
15:03:36 <Steven> Manu: We need to be compatible
15:04:32 <Zakim> -McCarron
15:05:24 <Steven> [ADJOURN]
15:05:38 <Zakim> -[MIT528]
15:05:47 <Steven> zakim, list attendees
15:05:56 <Zakim> -tinkster
15:06:00 <Zakim> -Ivan
15:06:06 <Zakim> -markbirbeck
15:06:10 <Zakim> -Knud
15:06:17 <Zakim> -Steven
15:06:22 <Zakim> As of this point the attendees have been tinkster, manu, Ivan, Steven, McCarron, [MIT528], Knud, BenjaminA, markbirbeck