<Larry_Masinter> scribenick: Larry_Masinter
Date: 02 Apr 2009
<Noah_Mendelsohn> zakim who is here?
John_Kemp regrets for the 16th
missing two people for next two weeks, will still schedule meetings
next scribe is Jonathan_Rees?
<Jonathan_Rees> next scribe is Jonathan_Rees.
dan: ftf minutes are ok [10:09]
Tim_Berners-Lee: minutes are ok
<John_Kemp_> +1 to f2f minutes
+1 to f2f minutes
resolution: publish f2f minutes publicly [10:11]
no one has reviewed minutes, put off one more week
Noah_Mendelsohn: reminder, 1.5 week ago email, hold discussion of group priorities in email
... hasn't been much discussion, though [10:13]
discussion of next F2F meeting -- can we talk about this now?
Noah_Mendelsohn: maybe TPAC in santa clara in november? [10:15]
<ht> ack danc [10:17]
<Zakim> Dan_Connolly, you wanted to offer KC/MCI and 2nd EDI and SFO
<raman> This is *a very * unproductive use of 9 busy individual's time
Tim_Berners-Lee week of 21st is open
<Jonathan_Rees> 9/21 bay area is ok with me
<John_Kemp_> I would like to check about any specific dates before committing [10:18]
note that week of 9/21
<Dan_Connolly> MIT/BOS is OK by me
proposal to host at MIT? ok?
<John_Kemp_> MIT/BOS is OK too
<Zakim> ht, you wanted to ask for later in Sept as opposed to earlier
<Dan_Connolly> ah. 21 Sep. folks did write it down. thanks.
<Dan_Connolly> 22-24 Sep in particular
<Tim_Berners-Lee> By the way, totally out of ordering the meeting, I could probably organize a TAG meeting in Centralish London, maybe at one of the REng, BCS, RS, etc ... so I would be intrested in people's interest in LHR as nearest airport. [10:25]
<Dan_Connolly> (I'm not insisting.)
<Dan_Connolly> +1 meet near LHR
<Ashok> I will be in Englad the previous week so I'm happy with a mtg nera LHR [10:27]
Noah_Mendelsohn: web application, HT and Raman --- publish as working draft [10:21]
web application state
Noah_Mendelsohn: worth changing URL to URI [10:22]
raman: not prepared to make the fine distinction
ht: global change only reasonable change
<Dan_Connolly> yes, global change. thanks.
<Tim_Berners-Lee> Please change all to URI
<Dan_Connolly> fragid are part of URIs, John_Kemp. the TAG fixed that a while ago
Noah_Mendelsohn: short answer is URI is the right term
Raman asked HT and HT agreed to make the global change
<Dan_Connolly> +1 publish subject to s/URL/URI/g
Noah_Mendelsohn: this is a working draft
ht: intends to send a plain text "status of this document"
<Dan_Connolly> (the SOTD is not up to us; it's up to the team; I'd rather our decision were orthogonal to the status section)
Noah_Mendelsohn: proposal: agree that we will publish that s/URL/URI/ and Henry will make a 'status of this document' change
ht notes that hash-in-uri will be the requested short name for the document URI
resolved: we will publish Web application state after substituting URL -> URI and after Henry adds a 'status of this document' section [10:29]
<trackbot> ISSUE-62 -- Uniform Access to Metadata -- OPEN
Jonathan_Rees: metadata is a much bigger topic than Issue-62
... prepared a survey of metadata, document is about metadata in general, but issue-62 is specifically about the protocol [10:30]
... will go through the document & say some things that are not in it
... as far as metadata formats go, W3C decided to use RDF for metadata 8-9 years ago. If you look up Metadata on W3C site, you find RDF [10:31]
... the term shows up in many groups, hard to figure out how to unify that
raman: is metadata on the web going in the same direction? [10:32]
Jonathan_Rees: found it useful to make a list about reasons to care about metadata: for search, browsing, decision making, .... listed in Jonathan_Rees's document
... very few kinds of metadata that apply to all documents
raman: there are 9 elements of dublin core
Larry_Masinter: dublin core applies
Jonathan_Rees: RDF is only a format, without a schema or ontology, haven't made any decisions [10:34]
... variations across schemas, syntax, where it's stored, and how you get it if it's not stored in the document. if it's not stored in the document, there are 10 listed and more around for accessing metadata
... not obvious where to go. Encourage people to establish vocabularies in RDF, but seems like they're going there anyway? [10:35]
<scribe> scribenick: Jonathan_Rees
Larry_Masinter: Work on metadata doesn't talk about process by which you find or create it [10:36]
... this is troubling, since you might end up with stuff you can't get at
... search enhancement, conflicts around the authority of the metadata [10:37]
... Three parts to XMP: 1. A profile of RDF for metadata. Adobe apps are content to make assertions about the media itself (the subject is always the document) [10:38]
<Tim_Berners-Lee> An implcit subject.
Larry_Masinter: 2. Set of schemas - applications to different kinds of media - partial independence from format (e.g. generic across image formats) [10:39]
... Some effort has gone into translations between metadata formats
scribe: Section3 of the spec: How to embed the attribute/value pairs into some popular media types
<Noah_Mendelsohn> I'm curious as to whether XMP is getting a lot of traction across the industry, or is it mainly used by Adobe software? [10:41]
scribe: In section 2, how to track provenance of objects as they get versioned and as they derive from other objects
... Main problem with metadata is that it's a set of opinions [10:42]
... Statements are not always intrinsic, added afterwards, subject to dispute
... Sometimes relative [?] to local policy, e.g. attribution [10:43]
<Larry_Masinter> ACTION: Masinter to send a summary [recorded in http://www.w3.org/2009/04/02-tagmem-minutes.html#action01]
<Larry_Masinter> scribenick: Larry_Masinter
Larry_Masinter: what is the TAG blog?
<Dan_Connolly> it's http://www.w3.org/blog/tag/
ashok: issue-62 is about access to metadata, how to find it and how to get hold of it [10:46]
... I don't think it is useful to us to worry about how to write it. It willbe hundreds of formats
... if we do, we should open a different issue on that
Jonathan_Rees: Eran's proposal allows you to subclass the protocol
... it would be nice to have a story
<Zakim> Tim_Berners-Lee, you wanted to mention that the work on the provenance of a document in terms of composition and editing very much of interest to the Media Standards Trust for example.
Tim_Berners-Lee: here we are, trying to get one interoperable web. To some extent we have to choose our battles. Every time we get a different proposal, there might be a problem. We could make a specific proposal
... XMP's use of RDF is quite ancient, it uses some older patterns, and the patterns for RDF [10:49]
... the fact that it doesn't have a subject stated was news to me
... does use things like alt and bag that aren't used as much [10:50]
<Jonathan_Rees> media standards trust?
Tim_Berners-Lee: the Media Standards Trust is a group started by a bunch of news feeds, for things on the web, for saying things about 'this photograph hasn't been retouched', to leave a trail
<Jonathan_Rees> http://www.mediastandardstrust.org/ [10:51]
http://www.adobe.com/devnet/xmp/ has specs. The compound object info is in the "Partners Guide to XMP"
<Zakim> Noah_Mendelsohn, you wanted to discuss goals
Noah_Mendelsohn: I'm hearing there are clearly some interesting issues for the web in this area, and potentially might be a good thing, might help the community settle on one [10:52]
... what should the TAG in particular do? "Zero-based budgeting" comes to mind: I don't think we should back into doing this
... is there anyone who would like to propose doing anything behind 'doing nothing' [10:53]
raman: happy for the tag to do something in this area as long as that someone takes on a concrete work item
Noah_Mendelsohn: someone did a good job of putting together an overview [10:54]
<Jonathan_Rees> Larry_Masinter: Unhappy about 62 being open without a broader story about schema/ontology and process
danc: RDFa vs GRDDL, Dan spent a whole day writing up a resolution of an issue, wasn't sure if anyone noticed it [10:56]
Dan_Connolly: record says we're all content, so we shouldn't be talking about it
<Dan_Connolly> 0047's subject is "Syntax and semantics for embedding RDF in XHTML" TAG issue addressed by RDFa and GRDDL (RDFinHTML-35) [10:58]
ashok: If you look at Jonathan's survey... (?) there is a very specific schema (?) different people have different requirements for putting in what we are calling metadata. Yes we can step up and work on some of them. To access them in their different formats is what ISSUE-62 is. I'm arguing that it is a reasonable issue to think about.
<Zakim> Jonathan_Rees, you wanted to note Eran is going ahead with LRDD regardless of what we say...
Jonathan_Rees: to answer Larry's comment on ISSUE-62, I wouldn't object to dropping ISSUE-62. Eran and others are going ahead with this protocol regardless of what we have to say. We can let it go, or individually respond. I'm happy to let it go. By design it doesn't address the question we're talking about here. [11:00]
<Zakim> Noah_Mendelsohn, you wanted to talk about scope of 62
Jonathan_Rees: the question about schema is different
<Jonathan_Rees> Access protocol, and details about what's accessed, are orthogonal. [11:01]
Noah_Mendelsohn: Larry's point is that it would hard to make progress on 62 without broadening the scope. Confusion that starts with the access method. I'm comfortable with a middle ground with leaving 62 open. I'm not convinced it's time for the TAG to take a new issue. If it becomes difficult to have a discussion about 62 without being explicit, we'll find that out. [11:02]
... we can help the community sort out some useful things without pulling on the whole ball of string.
<Jonathan_Rees> Larry_Masinter: Media annotations & POWDER are currently working in this space. [11:03]
<Jonathan_Rees> ... We should give priority to work that's ongoing in w3c. [11:04]
<Dan_Connolly> (I continue to wonder if the POWDER folks and the site-meta folks and the XRDS folks are talking to each other enough.)
<Zakim> Tim_Berners-Lee, you wanted to suggest separate issue metadata format and ontology separate fro 62.
Tim_Berners-Lee: not sure if we want to open a new issue?
Noah_Mendelsohn: open a new issue in background state?
ashok: what this issue would cover?
<Jonathan_Rees> -0.1 [11:06]
<Tim_Berners-Lee> A possible outcome of the issue would be to point to the existing RDF-based format
<Dan_Connolly> (new issue in the background doesn't seem cost-effective to me.)
Tim_Berners-Lee: Issue is standardization of format and ontologies for metadata, it's useful to have some standards and extensibility. A finding in this area might encourage people to use particular vocabularies, not invent new terms, etc.
<Jonathan_Rees> RDF is not very useful without agreement on schema (ontology). Agreement can be emergent or prescribed... Tim's preference may be emergent, mine might be prescribed... [11:07]
Tim_Berners-Lee: for me, issues are ways of organizing and categorizing things
volunteered to be shepherd [11:08]
<Dan_Connolly> (the volunteers to shepherd the issue address my concern. thanks.)
ashok: if you want to represent this .... in metadata, we will recommend this format?
... If you want to represent this type of metadata, then use this format?
Tim_Berners-Lee: could be also, for any metadata at all, you should use X. Would not constrain it. [11:09]
Noah: distinction between scope of issue vs. scope of findings in the area
... issue interoperability around formats for metadata on the web [11:10]
<Jonathan_Rees> Schema(ontology) recommendation is too big a job for the TAG, and possibly out of scope... [11:11]
Noah_Mendelsohn: the real question is: Issue-63? Representations of metadata on the web? [11:12]
<Dan_Connolly> (the question is complete, for my purposes, with just the issue name)
<ht> This is what I said at the f2f on issue 62: " There is a problem for which these people are conveging on a solution. It is of interest to the TAG as it seems architecturally sound. I want to keep this issue open and watch and help it. " [11:13]
propose issue: Metadata on the web: representation, schemas and their relationship to access methods
companion to issue 62
<Tim_Berners-Lee> (why their relation to access methods?)
<Tim_Berners-Lee> schemas should be ontologies
<Noah_Mendelsohn> TBL: Why the "relationship part"
<Noah_Mendelsohn> Larry: I'm concerned that you choose different representations for embedded vs. external
<Ashok> Prefer ontologies
tim: change schemas to ontologies [11:16]
+1 to changing it
<Noah_Mendelsohn> propose issue: Metadata on the web: representations & ontologies
<Ashok> +1 [11:17]
<Dan_Connolly> I'm +0 on metadata-framework and metadataArchitecture
<Tim_Berners-Lee> Metadata on the web: representations & ontologies
<Jonathan_Rees> +1 framework
<Dan_Connolly> (+0 is apache-speak for "I think that's OK, as long as somebody else does the work")
<John_Kemp_> +1 for framework too [11:18]
<Dan_Connolly> me thinks 2 words are sufficient now: any objections?
<Noah_Mendelsohn> propose: open issue metadataFramework-xx Metadata on the web: representations & ontologies
<Dan_Connolly> issue: metadataFramework [11:19]
<trackbot> Created ISSUE-63 - MetadataFramework ; please complete additional details at http://www.w3.org/2001/tag/group/track/issues/63/edit .
<Noah_Mendelsohn> Rank: background
<scribe> ACTION: larry to open MetadataFrmework issue & create rank of background and identify myself as shepherd [11:20] [recorded in http://www.w3.org/2009/04/02-tagmem-minutes.html#action02]
<Dan_Connolly> trackbot, status
<Noah_Mendelsohn> ACTION: larry to send email to www-tag announcing issue-63 [recorded in http://www.w3.org/2009/04/02-tagmem-minutes.html#action03]
<trackbot> Created ACTION-254 - Send email to www-tag announcing issue-63 [on Larry Masinter - due 2009-04-09].
<Noah_Mendelsohn> ack next [11:21]
<Noah_Mendelsohn> John: can you paste a link the note from Mark that you found so helpful? [11:23]
<raman> need to leave a few mins early, not saying bye so I dont break the flow
John: summarized email threads. Found a good summary from MNot. Basic issue is that creative commons have cc+ spec, with rel attribute with CURIE, doing work based on RDFa in XHTML spec, all documented in action
<trackbot> ACTION-240 -- John Kemp to read thread on RDFa, CURIEs and profile and summarize http://lists.w3.org/Archives/Public/www-tag/2009Feb/0295.html -- due 2009-03-21 -- PENDINGREVIEW
JohnK summarizes his email
JohnK: RDFa in XHTML allows two different syntax, neither is a QName [11:25]
... perhaps two obvious actions, ask XHTML what they are doing, or give Mark some direction about what he's doing in the Link draft
<Dan_Connolly> (is there a test case that captures this is-qname-allowed-here-or-not issue?)
JohnK: Third thing that is interesting that the CC+ spec will be transmitted in HTML 4, only well-regarded way is to use a profile, CC+ spec doesn't do that
Noah: HTML5 doesn't provide for profile attribute
<John_Kemp> ack Larry_Masinter [11:27]
<Zakim> Dan_Connolly, you wanted to note this is pretty clearly a request to re-open RDFinXHTML-35 Syntax and semantics for embedding RDF in XHTML http://www.w3.org/2001/tag/group/track/issues/35
Dan_Connolly: discussion on a mailing list that seems pertinent [11:29]
JohnK should contact Sam to see if he needs TAG help on this issue
<John_Kemp> mnot's note: http://lists.w3.org/Archives/Public/www-tag/2009Feb/0270.html [11:30]
discussion about next steps
John_Kemp to contact Sam Ruby to ask (a) how can the TAG be helpful (b) offer set up phone call involving Sam & 1 or 2 others [11:31]
<Dan_Connolly> ACTION: JohnK to contact Sam to ask (a) how can the TAG be helpful (b) offer set up phone call involving Sam & 1 or 2 others [recorded in http://www.w3.org/2009/04/02-tagmem-minutes.html#action04]
<trackbot> Created ACTION-255 - Contact Sam to ask (a) how can the TAG be helpful (b) offer set up phone call involving Sam & 1 or 2 others [on John Kemp - due 2009-04-09].
<Dan_Connolly> you did the hard part :) [11:32]
mnot expresses some concern that cc+ spec advocates something not consistent
Sam, Ben Adida, Mark Notthingham involved
Larry: want to make sure we get to versioning next time
<Dan_Connolly> ^ that's a start
<Noah_Mendelsohn> Hmm: Digg "URL shortener. Not convinced this is an entirely good thing: http://www.scripting.com/stories/2009/04/02/diggHasAUrlshortener.html [15:17]
<Noah_Mendelsohn> ..."nd now the Friends-of-Dave feed and the NY Times River all are running on the Digg shortener. They have, over the last few months been running on a variety of shorteners. Permalink to this paragraph" [15:18]
<Noah_Mendelsohn> Are URL shorteners a new trend I've managed to miss?
Larry_Masinter: yes [16:03]
... 5-10 years i think [16:04]