Chatlog 2011-11-03

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

13:38:19 <RRSAgent> RRSAgent has joined #rdfa
13:38:19 <RRSAgent> logging to http://www.w3.org/2011/11/03-rdfa-irc
13:38:21 <trackbot> RRSAgent, make logs world
13:38:23 <Zakim> Zakim has joined #rdfa
13:38:24 <trackbot> Zakim, this will be 7332
13:38:25 <Zakim> ok, trackbot; I see SW_RDFa()10:00AM scheduled to start in 22 minutes
13:38:27 <trackbot> Meeting: RDF Web Applications Working Group Teleconference
13:38:29 <trackbot> Date: 03 November 2011
13:44:34 <manu1> Guest: Niklas (lindstream) Lindström
13:44:42 <ShaneM> ShaneM has joined #rdfa
13:44:59 <ShaneM> ShaneM has left #rdfa
13:48:57 <ShaneM> ShaneM has joined #rdfa
13:54:53 <niklasl> niklasl has joined #rdfa
13:57:38 <tomayac> tomayac has joined #rdfa
13:59:03 <Zakim> SW_RDFa()10:00AM has now started
13:59:10 <Zakim> +??P4
13:59:13 <manu1> zakim, I am ??P4
13:59:13 <Zakim> +manu1; got it
13:59:51 <Steven> Steven has joined #rdfa
14:00:07 <ivan> zakim, dial ivan-voip
14:00:07 <Zakim> ok, ivan; the call is being made
14:00:09 <Zakim> +Ivan
14:00:31 <Zakim> +??P7
14:00:38 <niklasl> zakim, I am ??P7
14:00:38 <Zakim> +niklasl; got it
14:00:45 <Zakim> +Steven
14:01:10 <Zakim> +scor
14:01:11 <scor> scor has joined #rdfa
14:01:58 <Steven> zakim, who is on the call?
14:01:59 <Zakim> On the phone I see manu1, Ivan, niklasl, Steven, scor
14:02:07 <gregg> gregg has joined #rdfa
14:02:56 <gkellogg> gkellogg has joined #rdfa
14:03:18 <Zakim> +??P18
14:03:27 <ShaneM> zakim, ??P18 is ShaneM
14:03:27 <Zakim> +ShaneM; got it
14:03:48 <Zakim> +??P21
14:04:01 <gkellogg> zakim, ??P21 is gkellogg
14:04:01 <Zakim> +gkellogg; got it
14:04:43 <manu1> Agenda: http://lists.w3.org/Archives/Public/public-rdfa-wg/2011Nov/0008.html
14:05:13 <ShaneM> Scribe: ShaneM
14:06:10 <ShaneM> Topic: Supporting schema.org via RDFa Lite Changes
14:07:23 <ShaneM> Discussion at TPAC of @property and @src.  Didn't seem to be any controversy with that.
14:07:43 <manu1> q+ to ask about "@rel/@property being handled in a 'compatible' way?"
14:08:02 <ShaneM> Discussions about chaining.  Ben felt chaining is important and that some people just don't understand the requirement.  If @property is similar to @rel that's okay.
14:08:38 <manu1> ack manu1
14:08:38 <Zakim> manu1, you wanted to ask about "@rel/@property being handled in a 'compatible' way?"
14:08:02 <ShaneM> manu: I think it's clear that we want to support the changes that schema.org is asking us to make... the only question is the how, which has a solid proposal and the question of how much legacy RDFa 1.0 Markup is affected. I think that if we can address all of the schema.org issues and break less than 1% of existing RDFa markup out there... we should make the changes.
14:08:50 <ShaneM> Topic: Ivan's @typeof proposal
14:08:50 <ShaneM> http://lists.w3.org/Archives/Public/public-rdfa-wg/2011Nov/0001.html
14:08:50 <ShaneM> Brief discussion about Ivan's proposal.  There was some discussion of using @about as an object, but some didn't like that.  
14:09:26 <ivan> q+
14:10:24 <scor> Gregg: many ways to go about this - one possibility is to have @typeof apply to @resource
14:10:35 <gkellogg> <div property="foo" typeof="bar" resource="baz">
14:10:48 <gkellogg> <> :foo :baz
14:10:57 <gkellogg> :baz a :bar
14:11:45 <Zakim> + +1.404.978.aaaa
14:12:16 <tomayac> zakim, +1.404.978.aaaa is me
14:12:16 <Zakim> +tomayac; got it
14:12:19 <manu1> ack ivan
14:12:43 <manu1> q+ to say he's concerned about big changes like this.
14:12:56 <ShaneM> ivan: the whole thing is only if there is a property or a rel - without those the behavior does not change.  That use is relatively widespread.
14:14:09 <manu1> ack manu1
14:14:09 <Zakim> manu1, you wanted to say he's concerned about big changes like this.
14:14:40 <ivan> q+
14:14:45 <ShaneM> manu: This is a very large change when this started out as something small.  We dont have a clear grasp of how these changes might effect existing markup.
14:15:05 <ShaneM> ... it is fine to talk about them, but to make any changes, we need to understand the type of markup that is already out there. 
14:15:17 <ShaneM> ... there are some services that we could use to do a big crawl to see what is in use.
14:15:23 <Steven> +1
14:15:33 <manu1> ack ivan
14:15:34 <gkellogg> q+
14:15:36 <ShaneM> ... I am afraid that by creating 'consistency' with this one aspect of RDFa's use of @typeof will break more things than it fixes.
14:15:38 <ShaneM> +1
14:16:45 <ShaneM> ivan: I agree it is a concern.  I can't judge how big it is.  if we include the changes we are discussing in another draft, we will obviously need to ask for feedback from the community.
14:15:38 <ShaneM> Topic: "Magnetic @property" and "Chaining @property" changes
14:17:16 <ShaneM> Ivan: while playing with this I realized that if we have @property work this way it can make RDFa code simpler / easier (@property picks up @href).
14:17:34 <ShaneM> ... this is important.  The reason we are looking at this is because we have feedback that this is difficult to manage.
14:17:49 <manu1> q+ to take a position - Gregg's changes, but only if schema.org adopts RDFa and if it breaks "less than 1% of the markup out there"
14:18:18 <ShaneM> ... what gregg called chaining, which I would rather call the @typeof fix. If we look at the examples on schema.org, it is true that what you have to do in RDFa with @rel etc. is uglier and more complicated because of how @typeof works.
14:18:44 <manu1> ack gkellogg
14:18:59 <niklasl> q+
14:19:41 <ShaneM> Gregg: We need to be able to distinguish between 1.0 and 1.1 markup.  1.1 isn't a rec yet, so we aren't breaking things if we make changes to new behavior.
14:19:33 <manu1> Manu: I don't think anyone is suggesting that we're "breaking" RDFa 1.1 markup.
14:19:46 <manu1> Manu: The concern here is RDFa 1.0 backwards compatibility... we want to make sure that we don't break more than 1% of the markup that is already out there.
14:19:54 <ivan> q+
14:20:00 <manu1> ack ivan
14:20:09 <ShaneM> ... there is a concern about img and @src, but we have already broken that case.
14:20:23 <ShaneM> manu: I disagree.  We are not concerned about breaking 1.1 markup... it's not a REC yet.
14:21:08 <gkellogg> q+
14:21:15 <ShaneM> ... my concern is breaking 1.0 markup.  If we can change the rules so 1.0 markup generates relatively the same triples, that's okay.  Making changes that would do something unexpected to what is out there is a bad idea. Not knowing if it will destroy existing RDFa markup is even worse.
14:21:34 <ShaneM> ... saying that all we need to do is detect 1.0 markup is very difficult.  There is no announcement in much of the RDFa markup.
14:22:11 <scor> manu1: then when crawling the web for RDFa patterns, it should also include whether the version is specified or not
14:22:29 <ShaneM> ... I am perfectly fine making changes if we can prove that new markup idioms are not being used already in the wild.
14:23:05 <manu1> ack manu1
14:23:05 <Zakim> manu1, you wanted to take a position - Gregg's changes, but only if schema.org adopts RDFa and if it breaks "less than 1% of the markup out there"
14:23:07 <ShaneM> ... if no one is @typeof and @property on the same element, for example, then we can make that change.  If it is in use out there, that would be bad.
14:23:08 <manu1> ack niklasl
14:24:05 <niklasl> <div property="knows" typeof="Person">
14:24:06 <ShaneM> niklasl: I agree with what you say.  I hope we can make @typeof work.  It seems that the @typeof changes could be done with minor edge case issues.
14:25:00 <manu1> q?
14:25:15 <ShaneM> ivan: This example is a good one.  The way @typeof is defined in RDFa 1.0 is really broken, in my opinion.
14:25:34 <niklasl> (note that the above is same as <div typeof="Person" property="knows"> (in case attar order confuses our "intuition"…))
14:25:58 <manu1> I didn't have the same issue with @typeof, btw... (albeit anecdotal evidence with one data point...)
14:26:52 <manu1> ack gkellogg
14:27:12 <ShaneM> q+ to disagree about xmlns
14:30:00 <manu1> ShaneM: Two people have now indicated that if you use xmlns, that we're talking about RDFa 1.0 - that's just wrong. RDFa 1.0 uses xmlns: as a mechanism for defining prefix mappings... however, we do have xmlns in other XHTML documents.
14:30:11 <Steven> I agree strongly with Shane on this
14:30:14 <manu1> ShaneM: I don't think we can use it as a reliable switch.
14:30:15 <Steven> XForms too
14:31:11 <ShaneM> Topic: Web Crawl to Support RDFa Lite Changes
14:31:11 <ShaneM> manu: there seems to be agreement about the basic "magnetic @property" change.
14:31:54 <ShaneM> ... the other change - where @property magically chains if used with @typeof is less accepted, but we think it's workable. We need to start collecting data on these changes.
14:31:55 <manu1> http://blekko.com/webgrep is one possibility
14:32:07 <scor> q+ to talk about webgrep
14:32:09 <Steven> q+
14:32:14 <ShaneM> ack ShaneM
14:32:14 <Zakim> ShaneM, you wanted to disagree about xmlns
14:32:22 <scor> http://www.commoncrawl.org/ is another - better, but we need to write the query in Hadoop, etc.
14:32:28 <scor> q+ to talk about Common Crawl
14:32:56 <ShaneM> manu: there are some tools out there.  we could start collecting data to backup the changes we are talking about making.
14:32:57 <Steven> The problem with that approach is that it misses intranets. HTML5 used that technique to say that we don't need @rev
14:33:02 <ivan> q+
14:33:16 <ShaneM> ... we could also wait for google or yahoo to help us answer the question.
14:33:34 <manu1> ack scor
14:33:34 <Zakim> scor, you wanted to talk about webgrep and to talk about Common Crawl
14:34:08 <scor> webgrep
14:34:18 <scor> http://blekko.com/webgrep?page=view&id=8638947e26c2cf16e32d3487fc187ac5 tells us that we cannot use regexes there.
14:34:18 <ShaneM> scor: webgrep doesn't allow regular expressions as of now.
14:34:41 <ShaneM> ... it is possible to search for a tag, but you can't search for property and rel or whatever.
14:34:41 <tomayac> probably it was discussed before, but could Sindice be helpful for this task?
14:35:58 <ShaneM> scor: commoncrawl looks promising .  it isn't a big crawl, but it is free.  you need to have your own amazon instances to do the processing of the data.
14:38:08 <manu1> q+ to talk about what tests to do.
14:38:09 <ShaneM> ... we should on the wiki work out a regular expression we would use for inspecting the web.  Then when we do have access we can used the agreed upon regular expressions.
14:38:28 <niklasl> q+ on if we need regexps for this crawling though? aren't xpath or css selectors enough (or even better)?
14:39:13 <manu1> ack Steven
14:39:16 <manu1> ack ivan
14:39:28 <ShaneM> Steven: I dont like using webcrawls to backup arguments because they don't crawl intranets.  Also because it disenfranchises things that are not yet adopted.
14:39:54 <scor> q+ to talk about intranets
14:39:58 <ShaneM> ivan: we are not proposing using it to drop a feature.  we are using it to see if a change in behavior would break something.  that is a little bit different than the HTML5 / @rev case.
14:40:17 <ShaneM> ... the current situation is a little confusing.  Before we make a crawl we need a clear idea of what we're trying to find out.
14:41:00 <ShaneM> ... the existing wiki is sort of a hash of ideas.  Could we have a single page where all the @property related changes are spelled out in one place.
14:41:24 <ShaneM> ... we could use that page for the crawl, but also to publicize our proposals to schema.org and the community at large.
14:41:40 <ShaneM> ... we got a bunch of relatively positive reactions to RDFa Lite was published.  That's good.
14:42:08 <manu1> q?
14:42:10 <manu1> ack manu1
14:42:10 <Zakim> manu1, you wanted to talk about what tests to do.
14:42:11 <ShaneM> ... its possible that we will get similar feedback if we write this down concisely and promote it.
14:42:34 <ShaneM> ... I have tried to implement some of the proposed changes, but it gets confusing after a while.
14:42:58 <gkellogg> q+
14:42:59 <ShaneM> manu: I agree.  We need to capture the changes to the processing rules.  While we are doing the grep for the web we should look for everything we care about.
14:43:22 <ShaneM> ... there is no such thing as too much data.
14:44:13 <ShaneM> ivan: my problem is that there are combinations in use today on the web because 1.0 requires them.  a search wouldn't reveal if people like using the awkward usage.
14:44:28 <ShaneM> manu: I agree, but the purpose isn't to determine with people like the 1.0 usage.
14:44:44 <ShaneM> ivan: I am not sure how far we have to go in protecting backward compatibility.
14:44:56 <manu1> ack niklasl
14:44:56 <Zakim> niklasl, you wanted to comment on if we need regexps for this crawling though? aren't xpath or css selectors enough (or even better)?
14:44:59 <ShaneM> manu: I am not sure either.  But without data we don't even know what we are talking about.
14:45:22 <manu1> ack scor
14:45:22 <Zakim> scor, you wanted to talk about intranets
14:45:29 <ShaneM> niklasl: regexp might be hard to use.  if we found a service to suport xpath or selectors that might be better.
14:45:54 <ShaneM> scor: support for xpath might be hard to find.  regular expressions are widely supported.
14:46:36 <ShaneM> ... I agree with Ivan.  We should lay down all the ideas.  Next to them, record the regexps that would help identify if the pattern is already in use.
14:46:50 <manu1> +1 to what scor just said!
14:46:57 <ShaneM> ... w.r.t. using a crawl.  Crawl data can help us make decisions.  We shouldn't use it exclusively.
14:47:14 <ShaneM> gkellogg: We can use the wiki.  The data is separated now.  We can split it into its own page.
14:48:01 <ShaneM> ... w.r.t. regexp, we can probably do it with 2 regular expressions. Once we have the raw data, we could use xpath to do a finer grained analysis.
14:48:19 <ShaneM> ... I can write up the regular expressions.  I am not sure I know what is in Ivan's mind as to the @typeof changes.
14:48:36 <ShaneM> ... Ivan, can you do a wiki page on those?
14:49:14 <ShaneM> ivan: I think it would be better to have a single page.  If Gregg starts the page, then I can finish it (the timezones work out).
14:49:28 <manu1> q+ to talk about sequence
14:49:33 <manu1> ack gkellogg
14:50:18 <ivan> q+
14:50:28 <ivan> ack manu1 
14:50:28 <Zakim> manu1, you wanted to talk about sequence
14:51:59 <manu1> ack ivan
14:53:41 <Knud> Knud has joined #rdfa
14:54:56 <niklasl> q+
14:57:05 <manu1> ack niklasl
14:57:36 <manu1> q+ to get a set of actions down.
14:58:20 <ShaneM> Seems to be a sense that we agree to the @property changes but need to work out the technical details and ensure that we are not breaking backward compatibility too much.
14:58:43 <manu1> ack manu1
14:58:43 <Zakim> manu1, you wanted to get a set of actions down.
14:59:33 <manu1> ACTION: Gregg update the @property proposal - replace chaining discussion w/ new pattern - @typeof change.
14:59:33 <trackbot> Created ACTION-101 - Update the @property proposal - replace chaining discussion w/ new pattern - @typeof change. [on Gregg Kellogg - due 2011-11-10].
14:59:50 <manu1> ACTION: Ivan follow up on @typeof change that Gregg makes on the wiki.
14:59:51 <trackbot> Created ACTION-102 - Follow up on @typeof change that Gregg makes on the wiki. [on Ivan Herman - due 2011-11-10].
15:00:50 <manu1> ACTION: Gregg and Manu and Niklas to write regex patterns for RDFa "in the wild" markup.
15:00:50 <trackbot> Created ACTION-103 - And Manu and Niklas to write regex patterns for RDFa "in the wild" markup. [on Gregg Kellogg - due 2011-11-10].
15:02:55 <scor> gkellogg: the @property changes include both @property/@rel and @typeof right?
15:03:31 <gkellogg> scor, they just use @property, but we'll need to discuss the difference between @property and @rel
15:06:01 <manu1> q+ to end the call.
15:09:25 <ShaneM> Ivan will commit some changes to the RDFa core document that spell out how processing rules are evolving due to "Magnetic @property" direction.
15:09:26 <Zakim> -Steven
15:09:27 <Zakim> -manu1
15:09:27 <Zakim> -tomayac
15:09:29 <Zakim> -gkellogg
15:09:29 <Zakim> -scor
15:09:30 <Zakim> -Ivan
15:09:31 <Zakim> -niklasl
15:09:37 <Zakim> -ShaneM
15:09:38 <Zakim> SW_RDFa()10:00AM has ended
15:09:39 <Zakim> Attendees were manu1, Ivan, niklasl, Steven, scor, ShaneM, gkellogg, tomayac
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000200