Chatlog 2010-07-15

From RDFa Working Group Wiki
Jump to: navigation, search

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

13:53:46 <RRSAgent> RRSAgent has joined #rdfa
13:53:46 <RRSAgent> logging to http://www.w3.org/2010/07/15-rdfa-irc
13:57:00 <Knud> Knud has joined #rdfa
13:57:24 <manu> trackbot, prepare telecon
13:57:26 <trackbot> RRSAgent, make logs world
13:57:28 <trackbot> Zakim, this will be 7332
13:57:28 <Zakim> ok, trackbot; I see SW_RDFa()10:00AM scheduled to start in 3 minutes
13:57:29 <trackbot> Meeting: RDFa Working Group Teleconference
13:57:29 <trackbot> Date: 15 July 2010
13:57:47 <manu> Agenda: http://lists.w3.org/Archives/Public/public-rdfa-wg/2010Jul/0080.html
13:57:56 <manu> Chair: Manu
13:58:17 <manu> Regrets: Benjamin, BenA, MarkB
13:58:17 <manu> Present: Ivan Herman, Knud Möller, Manu Sporny, Shane McCarron, Steven, Toby Inkster
14:03:18 <manu> zakim, who is on the call?
14:03:18 <Zakim> On the phone I see [IPcaller], ShaneM, tinkster, Ivan, Steven, +3539149aaaa
14:03:42 <manu> scribenick: manu
14:03:45 <manu> Scribe: Manu
14:03:59 <manu> manu: Any additions to the agenda?
14:04:24 <manu> steven: Can we can talk about XHTML Basic and RDFa?
14:04:25 <ivan> q+
14:04:30 <manu> Topic: XHTML Basic and RDFa
14:04:42 <manu> steven: There has been discussion about the XHTML Basic checker 
14:04:58 <manu> steven: People have been concerned that checker doesn't allow use of RDFa in XHTML Basic.
14:05:15 <manu> steven: Issue is that the checker is based on W3C spec, so changing it isn't as trivial as one would want.
14:05:26 <manu> steven: We could provide an XHTML+RDFa DTD - that would resolve the issue.
14:06:22 <manu> toby: Does the XHTML+RDFa 1.1 rec include a DTD? Could XHTML+RDFa use the DTD for XHTML+RDFa 1.1?
14:06:24 <manu> ack ivan
14:06:39 <manu> ivan: Team discussion - just having just the DTD doesn't solve the problem of the mobile checker.
14:06:57 <manu> ivan: people that are behind mobile web checker are looking at re-using XHTML Basic plus RDFa DTD.
14:07:12 <manu> ivan: Steven and I will ask Shane to put in the 10 minutes that he will need to do to make this work.
14:07:27 <tinkster> toby: Can we include *two* DTDs in the XHTML+RDFa 1.1 Rec? i.e. an XHTML+RDFa DTD, and an XHTML Basic+RDFa DTD.
14:07:43 <manu> ivan: The XHTML+RDFa 1.1 in the appendix has a complete DTD for XHTML 1.1 - we could place another one in the same document for XHTML Basic + RDFa 1.1
14:07:53 <ShaneM> q+ to discuss DTDs
14:08:02 <manu> ivan: This is easy to do, we don't lose anything, and it gets RDFa validating in XHTML Basic.
14:08:22 <manu> shane: Yesterday on the PF WG, they want XHTML + ARIA + RDFa, and HTML4 + ARIA + RDFa
14:08:39 <manu> manu: Does RDFa WG need to be involved in this?
14:08:59 <manu> shane: No, PFWG will be dealing with these issues, just letting everyone know.'
14:09:12 <manu> manu: Great news that RDFa is being integrated by XHTML Basic community as well as ARIA folks.
14:10:11 <manu> Topic: ISSUE-26: Error Reporting Mechanism
14:10:51 <manu> manu: can't make much progress today - benjamin and mark aren't here.
14:10:55 <manu> manu: Ivan has updates for us.
14:11:10 <manu> Ivan: I've added language into RDFa Core, text for the simpler version of the error triples.
14:11:29 <manu> ivan: if we go with those, we still have to define error classes for classes other than the @profile one.
14:11:36 <manu> ivan: that's one thing I did.
14:11:53 <manu> ivan: The other thing is that the namespace document /ns/rdfa - there is an RDFa version of that document...
14:12:08 <manu> ivan: The RDFa version that I have has both the old terms and the new terms for error management.
14:12:24 <manu> ivan: We're waiting on Mark to either agree, or to prove us wrong as this being valid way to go forward.
14:14:43 <manu> toby: There are RDF libraries that process other languages.
14:14:52 <ivan> q+
14:15:09 <manu> ack shanem
14:15:09 <Zakim> ShaneM, you wanted to discuss DTDs
14:15:11 <manu> ack ivan
14:15:19 <manu> toby: like TURTLE, RDF/XML... those have other error reporting mechanisms - we shouldn't say that those mechanisms MUST use the Error Reporting mechanism that this specification defines.
14:15:38 <manu> ivan: I understand Toby's point, if I have the distiller and integrate it into RDFlib, the situation becomes more complicated.
14:15:40 <manu> q+
14:15:54 <manu> ivan: I would be happy to say that the generation of the processor graph is not required, it is SHOULD but not a MUST.
14:16:14 <manu> manu: Toby, would that address your concerns?
14:16:17 <manu> toby: yes
14:16:23 <manu> q-
14:19:46 <manu> manu: There is an argument to be made about having a common environment when dealing with RDFa. So regardless of whether you are using a Python, Ruby, C++, C# or Perl RDFa Processor, the same errors, in the same format, are required to be returned to you. People working with RDFa Processors won't have to learn a completely new error reporting mechanism if consistent error reporting is a requirement of RDFa Processors.
14:21:02 <manu> toby: If we have a SHOULD, if implementers implement the Error Reporting Mechanism, they MUST implement all errors.
14:21:59 <tinkster> i.e. "all or nothing"
14:22:27 <manu> PROPOSAL: The Error Reporting Mechanism in RDFa Core 1.1 is optional, but if an implementation includes it, the implementation MUST implement reporting of all errors.
14:22:37 <Steven_> +1
14:22:39 <ShaneM> +1
14:22:40 <manu> +1
14:22:40 <tinkster> +1
14:22:41 <ivan> +1
14:22:44 <Knud> +1
14:22:53 <manu> RESOLVED: The Error Reporting Mechanism in RDFa Core 1.1 is optional, but if an implementation includes it, the implementation MUST implement reporting of all errors.
14:23:16 <manu> Topic: ISSUE-20: Deep Processing of XMLLiteral
14:24:12 <ivan> q+
14:24:13 <manu> manu: Waiting on Mark 
14:24:16 <manu> ack ivan
14:24:31 <manu> manu: Why do we want to support Deep Processing of XMLLiterals? There doesn't seem to be a good use case.
14:25:21 <manu> ivan: I seem to remember that we had some discussion last December on the mailing list, it was decided that deep processing is required because this is what RDFa 1.0 defines.
14:25:28 <tinkster> no, RDFa 1.0 forbids it.
14:26:12 <manu> manu: Toby is correct, RDFa 1.0 forbids deep processing of XML Literals.
14:26:31 <manu> ivan: so, if RDFa 1.0 forbids it - why is this suddenly an issue for now?
14:26:41 <manu> tinkster: Stephane said that it's very useful.
14:27:15 <ShaneM> From RDFa 1.0: Once the triple has been created, if the [datatype] of the [current object literal] is  rdf:XMLLiteral, then the [recurse]  flag is set to false.
14:27:41 <manu> tinkster: It's useful when you're processing an XMLLiteral to specify the abstract, title, and body of an article. The abstract, title, and body /could/ have additional triples in them.
14:27:53 <manu> tinkster: but you also want to extract triples out of the data in the triple.
14:28:12 <ShaneM> if the datatype is NOT XMLLiteral - if it is some other XML String type, then it will still be deeply traversed.
14:28:18 <manu> ivan: This is also useful for RSS feeds - in case you want to keep the structure, but also want to express triples in the structure.
14:28:28 <manu> ivan: I'm a bit concerned about backwards compatibility issues...
14:28:39 <manu> ivan: we generate a new set of triples, which is fine... 
14:28:40 <tinkster> q+ to mention opt-in
14:29:10 <manu> ivan: So you have old RDFa content, that generates new triples.
14:29:35 <ShaneM> q+ to discuss tobys idea
14:29:44 <manu> ack tinkster
14:29:44 <Zakim> tinkster, you wanted to mention opt-in
14:29:53 <manu> tinkster: What about something that allows one to opt-in?
14:30:18 <manu> manu: what about datatype="rdfa:TransparentXmlLiteral"
14:30:32 <manu> shane: I think doing something like that is fine.
14:31:47 <manu> q+
14:31:48 <tinkster> what about rdfa:TransparentXmlLiteral rdfs:subclassOf rdf:XMLLiteral .
14:31:50 <manu> ack shanem
14:31:50 <Zakim> ShaneM, you wanted to discuss tobys idea
14:32:09 <manu> shane: can't  you follow your nose, and know its an XMLLiteral?
14:32:42 <manu> ivan: RDF environments don't work like that... defining a new datatype means a fairly complex thing... you have to define a new value-space, etc.
14:33:16 <manu> shane: Why can't we translate, in the RDFa Processor, rdfa:TransparentXmlLiteral to XMLLiteral and then continue processing?
14:33:42 <manu> ivan: I want to make sure we don't have the interpretation of a URL and then figure out what to do based on that URI.
14:33:52 <manu> ivan: It raises a lot of problems.
14:34:09 <manu> ivan: If we go that way, we may need to introduce another attribute for deep processing.
14:34:19 <manu> Ivan: if we do that, we should be very convinced that we need that.
14:35:27 <manu> manu: There are other technologies that could solve this issue - HTML5 data-*, the class attribute, etc.
14:35:40 <manu> ivan: This may be something for RDFa 2.0
14:36:33 <manu> tinkster: There may be another easier way to opt into it.
14:36:54 <manu> tinkster: if we can find a simple way to do it.
14:37:09 <ShaneM> I agree that it seems like a simple thing.... if there were a simple trigger I would be in favor of integrating it in 1.1
14:37:39 <manu> tinkster: let me have a think on it.
14:37:51 <manu> Topic: ISSUE-24: Case-sensitive terms in HTML5
14:38:00 <manu> manu: Shane has a proposal for this...
14:38:23 <manu> http://lists.w3.org/Archives/Public/public-rdfa-wg/2010Jul/0082.html
14:39:05 <Zakim> -ShaneM
14:39:18 <ivan> q+
14:40:15 <manu> shane: When referencing TERMs in http://www.w3.org/1999/xhtml/vocab - do a case insensitive comparison to figure out if we generate a triple.
14:40:39 <manu> ivan: I think we should combine this issue with other  issues: the default URI
14:41:11 <manu> ivan: there is a difference between RDFa 1.0 and what we have right now - RDFa 1.0 has a fixed set of terms.
14:41:23 <manu> ivan: That set is listed in the document and those are the terms in the default vocab.
14:41:35 <manu> ivan: In RDFa 1.1, we don't have a list of terms.
14:41:41 <manu> q+ to discuss default URI
14:41:45 <manu> ack ivan
14:41:47 <Zakim> +McCarron
14:42:10 <ShaneM> zakim, mccarron is ShaneM
14:42:10 <Zakim> +ShaneM; got it
14:43:02 <manu> manu: We do specify the terms in XHTML+RDFa and HTML+RDFa
14:43:26 <tinkster> In RDFa 1.1, rel="quux" maps to xhv vocab; in RDFa 1.0 it maps to nothing.
14:43:57 <manu> shane: RDFa 1.0 enumerated these terms.
14:44:07 <manu> shane: I objected at  the time, and still now, because the list is not extensible.
14:44:21 <manu> shane: When the new spec got produced, it just got included.
14:44:43 <ShaneM> Note that the values defined in this section may be removed from this document and placed in an external  'RDFa Profile' so that they can be maintained independent of the specification.
14:45:02 <manu> ivan: The whole issue of lower-case and upper-case terms are related... that's all I was saying.
14:46:30 <ivan> q+
14:46:44 <manu> ack [IP
14:46:44 <Zakim> [IPcaller], you wanted to discuss default URI
14:46:48 <manu> ack ivan
14:46:50 <Steven_> zakim, [IP is Manu
14:46:50 <Zakim> +Manu; got it
14:46:57 <manu> zakim, who is on the call?
14:46:57 <Zakim> On the phone I see Manu, tinkster, Ivan, Steven, +3539149aaaa, ShaneM
14:47:05 <Steven_> q?
14:47:40 <manu> ivan: We should define the terms in the document - it's not extensible.
14:47:44 <ShaneM>  q+ about html5
14:47:45 <tinkster> Zakim, aaaa is probably Knud
14:47:46 <Zakim> +Knud?; got it
14:48:12 <ShaneM> ack about
14:48:13 <manu> ivan: keeping the terms in the document as it is now, would work.
14:48:18 <ShaneM> ack html5
14:48:23 <manu> ivan: We should remove concept of default vocabulary from document
14:48:24 <ShaneM> q+ to discuss html5 terms
14:48:34 <manu> ack shanem
14:48:34 <Zakim> ShaneM, you wanted to discuss html5 terms
14:50:55 <ivan> q+
14:51:10 <manu> shane: Manu, you had said that there were IETF and HTML5 rammifications.
14:51:20 <manu> manu: Yeah, so there are two separate issues here.
14:51:28 <manu> manu: One of them is that there is no longer a default vocabulary specified in HTML5.
14:51:32 <manu> manu: The other is that there are multiple LinkType registries now - one on the WHATWG Wiki, the other in the IETF, and another still in the XHTML/HTML vocab document. It would be good if we could harmonize all of these.
14:51:37 <manu> ivan: We could say that there is a profile file at a well-defined place that contains the HTML-related terms.
14:52:28 <manu> ivan: We could get into this type of issue: XML Tools get a DTD from our servers - but tools can cache DTDs for further processing.
14:52:48 <manu> ivan: So, we could publish a profile document for TERMS, but we could say that processors SHOULD cache the profile.
14:53:02 <manu> ivan: Authors are not required to use the profile for HTML and XHTML.
14:53:10 <manu> ivan: That's one way of doing things in a more flexible manner.
14:53:18 <manu> ivan: Sounds dangerous, but I don't see any other way.
14:53:36 <manu> shane: I wouldn't mind having a default profile specified.
14:53:48 <manu> shane: We already encourage processors to cache profiles.
14:55:13 <manu> toby: keywords could conceptually be a profile, but one that's hard-coded into XHTML+RDFa processors.
14:56:09 <manu> shane: I'd be okay with it being hardcoded... but in XHTML modularization, we make the commitment, we won't change a module without changing its URI
14:56:16 <manu> q+
14:56:26 <manu> ack ivan
14:58:27 <manu> manu: Why aren't we depending on RDFa Profile mechanism? We should depend on it.
14:58:42 <manu> shane: Going back to the issue at hand.
14:58:57 <manu> shane: We have to say that generated terms are case-sensitive
15:00:29 <manu> q+ to end the telco
15:00:52 <Zakim> -Steven
15:01:39 <tinkster> I think including a default @vocab, even for HTML and XHTML will result in lots of junk triples. We have to share @rel/@rev space with microformats, JS libraries, etc.
15:02:44 <manu> PROPOSAL: For terms in the default vocabulary, comparison should be performed in a case-insensitive manner when determining whether or not to generate a triple.
15:03:04 <manu> PROPOSAL: For terms in the XHTML/HTML vocabulary, comparison should be performed in a case-insensitive manner when determining whether or not to generate a triple.
15:03:13 <tinkster> +1
15:03:16 <ShaneM> -1
15:03:30 <manu> shane: I'm not comfortable with the state of the current discussion - we should slow down and discuss this more.
15:03:41 <tinkster> (this is already the case in RDFa 1.0, as an errata)
15:04:13 <Zakim> -ShaneM
15:04:16 <ivan> zakim, drop me
15:04:16 <Zakim> Ivan is being disconnected
15:04:18 <Zakim> -tinkster
15:04:18 <Zakim> -Ivan
15:04:18 <Zakim> -Manu
15:04:19 <Zakim> -Knud?
15:04:19 <Zakim> SW_RDFa()10:00AM has ended
15:04:21 <Zakim> Attendees were [IPcaller], ShaneM, tinkster, Ivan, Steven, +3539149aaaa, Manu, Knud?
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000219