Warning:
This wiki has been archived and is now read-only.

Chatlog 2011-07-28

From RDFa Working Group Wiki
Jump to: navigation, search

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

13:56:59 <RRSAgent> RRSAgent has joined #rdfa
13:56:59 <RRSAgent> logging to http://www.w3.org/2011/07/28-rdfa-irc
13:57:01 <trackbot> RRSAgent, make logs world
13:57:01 <Zakim> Zakim has joined #rdfa
13:57:03 <trackbot> Zakim, this will be 7332
13:57:04 <trackbot> Meeting: RDF Web Applications Working Group Teleconference
13:57:04 <trackbot> Date: 28 July 2011
13:57:04 <Zakim> ok, trackbot; I see SW_RDFa()10:00AM scheduled to start in 3 minutes
13:57:06 <manu1> Chair: Manu
13:57:06 <manu1> Guest: Stéphane (scor) Corlosquet
13:57:06 <manu1> Guest: Henri (bergie) Bergius
13:57:06 <manu1> Guest: Niklas (lindstream) Lindström
13:57:50 <manu1> Agenda: http://lists.w3.org/Archives/Public/public-rdfa-wg/2011Jul/0058.html
13:57:58 <Zakim> SW_RDFa()10:00AM has now started
13:58:05 <Zakim> +??P14
13:58:22 <lindstream> zakim, I am ??P14
13:58:22 <Zakim> +lindstream; got it
13:58:30 <Zakim> +??P17
13:58:33 <manu1> zakim, I am ??P17
13:58:33 <Zakim> +manu1; got it
13:59:44 <Zakim> +??P19
13:59:53 <gkellogg> zakim, i am ??P19
13:59:54 <Zakim> +gkellogg; got it
14:00:24 <Zakim> +OpenLink_Software
14:00:31 <MacTed> Zakim, OpenLink_Software is temporarily me
14:00:33 <Zakim> +MacTed; got it
14:00:35 <MacTed> Zakim, mute me
14:00:37 <Zakim> MacTed should now be muted
14:00:44 <MacTed> RRSAgent, draft minutes
14:00:44 <RRSAgent> I have made the request to generate http://www.w3.org/2011/07/28-rdfa-minutes.html MacTed
14:00:57 <tomayac> tomayac has joined #rdfa
14:00:57 <MacTed> RRSAgent, make logs public
14:02:19 <Zakim> + +1.781.866.aaaa
14:02:19 <Knud> Knud has joined #rdfa
14:02:36 <Zakim> + +3539149aabb
14:02:43 <Zakim> + +358.405.25aacc
14:02:50 <MacTed> Zakim, unmute me
14:02:50 <Zakim> MacTed should no longer be muted
14:02:52 <Knud> zakim, i am aabb
14:02:52 <Zakim> +Knud; got it
14:02:57 <bergie> zakim, I am aacc
14:02:57 <Zakim> +bergie; got it
14:03:01 <bergie> zakim, mute me
14:03:01 <Zakim> bergie should now be muted
14:03:37 <Knud> zakim, mute me
14:03:37 <Zakim> Knud should now be muted
14:03:37 <scor> scor has joined #rdfa
14:04:02 <MacTed> Zakim, who's here?
14:04:02 <Zakim> On the phone I see lindstream, manu1, gkellogg, MacTed, +1.781.866.aaaa, Knud (muted), bergie (muted)
14:04:04 <Zakim> On IRC I see scor, Knud, tomayac, Zakim, RRSAgent, lindstream, MacTed, danbri, bergie, gkellogg, manu1, manu, trackbot
14:04:16 <scor> zakim, I'm aaaa
14:04:16 <Zakim> I don't understand 'I'm aaaa', scor
14:04:24 <MacTed> Zakim, aaaa is scor
14:04:27 <Zakim> +scor; got it
14:04:29 <scor> zakim, aaaa is me
14:04:31 <Zakim> sorry, scor, I do not recognize a party named 'aaaa'
14:05:00 <manu1> zakim, who is on the call?
14:05:00 <Zakim> On the phone I see lindstream, manu1, gkellogg, MacTed, scor, Knud (muted), bergie (muted)
14:05:06 <MacTed> Zakim, mute me
14:05:09 <Zakim> MacTed should now be muted
14:05:12 <Zakim> +tomayac
14:06:42 <manu1> scribenick: tomayac
14:07:28 <tomayac> manu1: Do we need to discuss http://structured-data.org/ ?
14:07:41 <tomayac> gkellogg: no news, continous work, but not ready for release yet
14:08:20 <tomayac> manu1: Tantek Çelik from Microformats community and Philip Jägenstedt and Henri Sivonen from Microdata community have been added to the structured-data code repository as admins. We are attempting to build bridges - hopefully this is a good first step.
14:09:17 <manu1> Topic: Proxy vocabularies via @vocab
14:09:43 <lindstream> https://gist.github.com/1092350
14:09:48 <manu1> http://lists.w3.org/Archives/Public/public-rdfa-wg/2011Jul/0048.html
14:10:02 <Zakim> +??P48
14:10:08 <ShaneM> ShaneM has joined #rdfa
14:10:16 <ShaneM> zakim, who is here?
14:10:16 <Zakim> On the phone I see lindstream, manu1, gkellogg, MacTed (muted), scor, Knud (muted), bergie (muted), tomayac, ??P48
14:10:19 <Zakim> On IRC I see ShaneM, scor, Knud, tomayac, Zakim, RRSAgent, lindstream, MacTed, danbri, bergie, gkellogg, manu1, manu, trackbot
14:10:22 <ShaneM> zakim, I am ??P48
14:10:22 <Zakim> +ShaneM; got it
14:11:00 <tomayac> Niklas talks us through https://gist.github.com/1092350 - Basically a Proxy Vocabulary would allow people to define uber-vocabularies, like http://schema.org/ , but also ensure that there is a clear mechanism of mapping those vocabularies to dc, foaf, sioc or other more widely used vocabularies.
14:11:56 <tomayac> manu1: this almost looks like a re-invention of rdfa profiles. not saying it's a bad thing, just want to understand why we have things like map:ProxyTerm and map:ProxyClass.
14:12:06 <Knud> should this be part of the RDFa spec, or a separate spec (a W3C note?) that do we reference?
14:12:12 <tomayac> lindstream: it grew from that (RDFa Profiles). similarity is intentional.
14:13:20 <Knud> q+
14:13:37 <tomayac> manu1: Proxy Vocabularies are a post-processing step
14:14:04 <tomayac> manu1: wondering why not just using rdfs instead of map?
14:14:21 <gkellogg> +1 to talk about diff with rdfs
14:14:51 <gkellogg> q+ to talk about diff with rdfs
14:15:47 <Knud> zakim, unmute me
14:15:47 <Zakim> Knud should no longer be muted
14:15:48 <manu1> ack Knud
14:16:08 <tomayac> knud: do you think this should be a part of rdfa, or a best practice?
14:16:28 <lindstream> q+
14:16:41 <tomayac> lindstream: i don't see this as a part of rdfa
14:16:45 <manu1> ack gkellogg
14:16:45 <Zakim> gkellogg, you wanted to talk about diff with rdfs
14:17:00 <tomayac> gkellogg: this can be more generic than just rdfa
14:17:12 <tomayac> gkellogg: it includes some of the rdf entailment rules
14:17:23 <tomayac> gkellogg: you don't need something like owl:sameas which requires multiple iterations at times and is NP Complete.
14:18:03 <tomayac> gkellogg: Proxy Vocabularies are simpler - we probably do want something normative about it
14:18:05 <manu1> ack lindstream
14:18:33 <tomayac> lindstream: it would be good to publish a note apart from the rdfa spec
14:18:45 <tomayac> lindstream: there might be a need for this pattern elsewhere
14:19:09 <manu1> q+ to respond directly to Knud's question
14:19:28 <tomayac> lindstream: not sure if it's better than mixing vocabularies, but imho it's very valuable
14:19:28 <Knud> q+
14:19:46 <manu1> ack manu1
14:19:46 <Zakim> manu1, you wanted to respond directly to Knud's question
14:20:05 <tomayac> manu1: we should say something normative about it
14:20:34 <tomayac> manu1: mapping might be done differently by google, facebook, etc.
14:21:02 <tomayac> manu1: there's an opportunity to be specific on how we expect follow your nose to work.
14:21:04 <manu1> ack knud
14:21:21 <tomayac> knud: not yet completely convinced
14:21:47 <tomayac> knud: a stand-alone doc might be more useful
14:22:24 <tomayac> knud: facebook and google don't seem to want people to mix vocabularies, also microdata people think uber-vocabularies would be easier
14:22:58 <manu1> q+ 
14:23:08 <tomayac> knud: being able to mix vocabularies was our main motiviation to introduce profiles
14:23:09 <manu1> ack manu1
14:23:29 <tomayac> knud: but it seems that the big players don't want / need this, though
14:23:52 <tomayac> manu: I'm concerned about the W3C process
14:24:23 <lindstream> q+
14:24:25 <tomayac> manu1: pushing this separately as a REC is out of charter
14:24:38 <tomayac> manu1: we can put it in the rdfa core spec, but there are also issues with doing that - we may be too specific to RDFa - Proxy Vocabularies are a more general solution.
14:24:53 <tomayac> manu1: we can always publish it as a note and people can either ignore or use it
14:25:07 <tomayac> lindstream: i didn't ask to be normative about this
14:25:21 <manu1> ack lindstream
14:25:58 <gkellogg> q+
14:26:00 <Knud> zakim, mute me
14:26:00 <Zakim> Knud should now be muted
14:26:03 <manu1> ack gkellogg
14:26:05 <tomayac> manu1: I would be fine with publishing as a note
14:26:34 <ShaneM> q+ to ask what problem are we trying to solve?
14:26:57 <tomayac> gkellogg: doing it in a non-normative way might result in everyone just doing it their way
14:27:03 <manu1> ack ShaneM
14:27:04 <Zakim> ShaneM, you wanted to ask what problem are we trying to solve?
14:27:05 <tomayac> gkellogg: might lead to confusion
14:27:31 <tomayac> shanem: we created this WG with a set of use cases and a plan to address them
14:27:57 <tomayac> shanem: if we come to the end without addressing the use cases, we should just say we don't address them
14:28:37 <tomayac> manu1: we were very concerned about vocabulary mixing
14:28:46 <MacTed> +1
14:28:54 <tomayac> shanem: if we come out of this without a way to do it, i feel like we have not done our job
14:29:26 <MacTed> Zakim, unmute me
14:29:26 <Zakim> MacTed should no longer be muted
14:29:39 <tomayac> manu1: does not be normative mean you (OpenLink) wouldn't put it in Virtuouso
14:30:37 <Knud> vocabulary mixing can always be done with @prefix, right?
14:30:55 <tomayac> macted: not necessarily, @prefix doesn't allow term re-mixing?
14:31:22 <tomayac> manu1: would proxy vocabularies be a good replacement?
14:32:02 <tomayac> shanem: i would agree
14:32:12 <tomayac> macted: not sure what problem is being resolved
14:33:18 <tomayac> manu1: option 1 = modify prefix, 2 = use vocab, but limits to one vocabulary
14:33:46 <tomayac> manu1: proxy vocabulary allows for mixing via a big über vocabulary
14:34:10 <ShaneM> q+ to ask for a clarification on @vocab 
14:34:24 <manu1> ack shanem
14:34:24 <Zakim> ShaneM, you wanted to ask for a clarification on @vocab
14:34:45 <tomayac> shanem: if I'm a semantic data processing engine, and I see a term called 'foo', I'm gonna follow my nose to the profile document
14:35:14 <lindstream> q+
14:35:16 <tomayac> shanem: when I do that, I see that 'foo' maps to 'foaf:name'. Shouldn't a hardcore semantic data processing engine be doing that anyway?
14:35:32 <manu1> ack lindstream
14:35:34 <tomayac> manu1: in theory this is how it works, in practice - I don't think there are many people doing that.
14:35:52 <tomayac> lindstream: it is true that we already have all the semantic mechanisms there
14:36:11 <manu1> q+ to add normative statements to @vocab
14:36:21 <tomayac> lindstream: it is more or less putting a focus on how to do it. Proxy Vocabularies are a middle-ground between hardcore entailment/reasoning and dumb processing. If we want this stuff to work well in browsers, we can't expect an NP Complete solution to get traction.
14:37:18 <gkellogg> q+
14:37:30 <tomayac> shanem: if the semweb isn't doing this yet, then probably they never will. who needs the hint? 
14:37:38 <tomayac> lindstream: the non-hardcore engines
14:37:45 <tomayac> shanem: but they don't seem to care
14:37:51 <tomayac> lindstream: fair enough...
14:37:52 <manu1> ack manu1
14:37:52 <Zakim> manu1, you wanted to add normative statements to @vocab
14:38:15 <tomayac> manu1: what if we said normatively that you should use proxy vocabularies if you plan to do an uber vocabulary?
14:38:28 <lindstream> q+
14:39:07 <tomayac> manu1: we shouldn't have to say this
14:39:23 <manu1> ack gkellogg
14:39:24 <tomayac> manu1: we expect people that use vocab to make sure that people can follow their nose
14:39:43 <tomayac> gkellogg: we have to remember that there's a difference between processing and reasoning
14:40:07 <tomayac> gkellogg: if we ever want to have browser vendors to do stuff with this, reasoning won't be the way
14:40:08 <ShaneM> text from the role spec about vocabluaries: It is possible to define additional role values. Such values must be defined in their own vocabulary. The URI associated with that vocabulary should resolve to a resource that allows for the machine and human discovery of the definition of the roles in the vocabulary. One format that achieves this is the RDFa Profile as defined in [RDFA-CORE].
14:40:22 <manu1> ack lindstream
14:40:52 <tomayac> lindstream: since vocab is used to turn the term into a uri, you should be able to get info on what it is
14:41:26 <ShaneM> I am happy to add text to RDFa Core that indicates the URI for @vocab must resolve and SHOULD resolve to a resource that ... (see above)
14:41:56 <tomayac> lindstream: hardcore semweb engines follow their noses, others don't
14:42:26 <tomayac> manu1: seems like we have come full circle here, this stuff should just work in an "ideal world" - however, we are not operating in that and some guidance would be good.
14:43:55 <tomayac> shanem: if the community is unable to figure out follow your nose reliably, the semantic web community has failed miserably in explaining it correctly. How is this problem? We're just writing a small spec on top of the sem web stack - isn't this the responsibility of the larger semweb community?
14:44:16 <lindstream> q+
14:45:13 <ShaneM> Ted: put it in the primer
14:45:27 <tomayac> shanem: we are now endlessly complicating
14:46:29 <Knud> isn't there a W3C document somewhere called "the self-describing web"?
14:46:33 <ShaneM> the sem web coord group may surprise you and point you at an existing spec that describes follow your nose
14:46:50 <ShaneM> http://www.w3.org/2001/tag/doc/selfDescribingDocuments.html
14:47:06 <Knud> yes, that
14:48:09 <ShaneM> q+ to clarify whether we want language in rdfa core that indicates that a vocab should include an RDFa definition of the vocab
14:48:21 <manu1> ack lindstream
14:48:51 <MacTed> it just seems that we're doing s/@profile/@proxyvocab/g
14:49:05 <MacTed> not really changing what's happening -- just the term for it
14:49:07 <tomayac> manu1: the semweb coordination group might be a good candidate for publishing this doc
14:49:24 <manu1> ACTION: Manu to raise @profile and Proxy Vocabulary issue with SWCG.
14:49:25 <trackbot> Created ACTION-88 - Raise @profile and Proxy Vocabulary issue with SWCG. [on Manu Sporny - due 2011-08-04].
14:49:50 <ShaneM> q?
14:49:54 <manu1> ack shanem
14:49:54 <Zakim> ShaneM, you wanted to clarify whether we want language in rdfa core that indicates that a vocab should include an RDFa definition of the vocab
14:50:20 <tomayac> shanem: do you want me to put that sentence (above) in the spec?
14:50:22 <Knud> RDFa or other RDF?
14:50:30 <tomayac> shanem: objections? none
14:50:33 <Knud> zakim, unmute me
14:50:33 <Zakim> Knud should no longer be muted
14:50:53 <ShaneM> The URI associated with that vocabulary should resolve to a resource that allows for the machine and human discovery of the definition of the roles in the vocabulary. One format that achieves this is the RDFa Profile as defined in [RDFA-CORE].
14:51:47 <lindstream> q+
14:51:50 <ShaneM> RDFa Profiles are collections of terms, prefix mappings, and/or default  vocabulary declarations.  A profile is either intrinsically  known to  the parser, or it is loaded as an external document and processed. These documents must be defined in an approved RDFa Host Language (currently XML+RDFa and XHTML+RDFa [XHTML-RDFA]).  They may also be defined in other formats (e.g., RDF/XML [RDF-SYNTAX-GRAMMAR],  or Turtle [TURTLE]). RDFa Profiles are referenced via @profile
14:52:00 <manu1> ack lindstream
14:52:17 <tomayac> lindstream: i agree. it is just a uri 
14:52:39 <tomayac> Topic: Removing @profile
14:52:39 <tomayac> manu1: are we ok with removing profile entirely from rdfa core?
14:53:02 <tomayac> manu1: with the assumption that the vocabulary stuff will be cleared
14:53:09 <MacTed> q+
14:53:28 <manu1> ack macted
14:53:58 <tomayac> macted: we don't seem to have a clear understanding of what the problem is
14:54:03 <Knud> q+
14:54:04 <lindstream> q+
14:54:08 <manu1> ack knud
14:54:23 <tomayac> knud: the problem with profile is that rdfa core requires it to be resolved
14:54:30 <tomayac> knud: what if it can't ?
14:54:45 <tomayac> knud: vocab does not have to be resolved, it can be, though
14:55:19 <manu1> ack lindstream
14:55:32 <tomayac> manu1: to clarify: RDFa Profiles via @profile is a 'must pre-process' mechanism, Proxy Vocabularies via @vocab is a 'may post-process' mechanism
14:55:45 <ShaneM> not MAY pre-process.  CAN post-process.
14:56:02 <ShaneM> And, FWIW, it was ALWAYS this way (@vocab)
14:57:08 <gkellogg> Without giving some guidance on /how/ to process @vocab, nothing will happen. We should have text indicating /what/ should be in the doc pointed to by @vocab with simple processing rules.
14:58:12 <gkellogg> A potential issue with multiple @vocab definitions is that each sub-graph needs to be processed using different rules.
14:59:27 <tomayac> manu1: we would have almost the same functionality covered by @profile and @vocab if we did 'may post-process' for both of them. We only need one solution. Both allow vocabulary mixing.
14:59:40 <lindstream> q+
15:00:15 <tomayac> shanem: you can declare prefix mappings, this wouldn't be possible in a post-process world
15:00:15 <tomayac> manu1: That is true - we would have to ensure that @prefix allowed one to declare terms.
15:01:19 <manu1> ack lindstream
15:02:12 <manu1> PROPOSAL: Remove @profile feature from the RDFa Core specification with the caveat that the issue will be re-opened if ePub, OpenDoc, IPTC, or other communities have an issue with its removal.
15:02:37 <Knud> +1
15:02:39 <tomayac> +1
15:02:41 <manu1> +1
15:02:41 <lindstream> +1
15:02:42 <gkellogg> +1
15:02:49 <MacTed> +1
15:02:52 <scor> +1 for removal
15:02:57 <bergie> +1
15:03:02 <ShaneM> +1
15:03:10 <manu1> RESOLVED: Remove @profile feature from the RDFa Core specification with the caveat that the issue will be re-opened if ePub, OpenDoc, IPTC, or other communities have an issue with its removal.
15:03:23 <Zakim> -bergie
15:03:25 <Zakim> -MacTed
15:03:27 <Zakim> -gkellogg
15:03:28 <Zakim> -Knud
15:03:33 <Zakim> -tomayac
15:03:37 <Zakim> -scor
15:10:03 <tomayac> rrsagent, draft minutes
15:10:03 <RRSAgent> I have made the request to generate http://www.w3.org/2011/07/28-rdfa-minutes.html tomayac
15:11:14 <tomayac> zakim, bye
15:11:14 <Zakim> leaving.  As of this point the attendees were lindstream, manu1, gkellogg, MacTed, +1.781.866.aaaa, +3539149aabb, +358.405.25aacc, Knud, bergie, scor, tomayac, ShaneM
15:11:14 <Zakim> Zakim has left #rdfa
15:11:22 <tomayac> rrsagent, draft minutes
15:11:22 <RRSAgent> I have made the request to generate http://www.w3.org/2011/07/28-rdfa-minutes.html tomayac
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000267