13 Dec 2006

See also: Agenda, Mon 11 Dec, Tue 12 Dec, IRC log


Tim, Norm, Ed, Henry, Vincent, Noah, Dan, Dave[approx 10:00-], TV[by phone after noon]


Semantic Web

Vincent: Norm sent a draft document.

<timbl> http://lists.w3.org/Archives/Member/tag/2006Dec/att-0034/SEMWEB.html

Norm: In fairness, TimBl really wrote most of it, I just did a little light editing

<DanC_lap> Data on the Web

<DanC_lap> TimBL: never mind the bits under "Fodder"

TimBL: it's a bit long winded, story wise.
... It's talking about the way the world should be, these aren't the sorts of things you get off-the-shelf today.
... What's missing are some pictures

TimBL draws: Semantic Web diagram (or photo by VQ)

Noah: The only thing about this that seems web-related is the web service.
... You can tell the standards story pretty well.
... I would have expected the disaster story to include communication with other manufacturers who might also have data in RDF.

TimBL: This tells the story of enterprise integration

Noah: Everyone is selling web services software for enterprise integration by recognizing that corporate boundaries have become more fluid.

TimBL: That's an important story to tell, but early adopters can't rely on the network effect.
... So should we tell this story and then the network effect story.
... We could extend the story but I'm not sure we can expect readers to make the leap.

Norm: We could tell a broader story, but I agree with Henry that keeping the story small is valuable.

Henry: One of the huge strengths of Feinman's Q.E.D. is that it says at the very beginning that everything in the book is true.

<DanC_lap> QED: The Strange Theory of Light and Matter

Henry: I want to be sure I know where the story we're telling could be done tomorrow and where it's vision.
... So there really is a web service that will turn a street address into a latitude and longitude.

DanC: Yes, there are two.

Henry: What about a data browser?

DanC: No, that's a little novel. There are a couple, but they're not widely used.

TimBL: I think we should include screen shots.

Norm: We could include examples of the transitions to RDF.

Henry: This fits very well with the observation that you don't sell technology, you sell problems.
... These are data integration problems.

TimBL: One other answer is to just put everything in one big database.
... For the next story, that solution doesn't work at all.

Henry: Even in this case, you have to get agreement across the company for what the database architecture is going to be. That's very hard.
... The problems here are more tractable because the boundaries are more fluid.

TimBL: The fact that there are discrepancies between Joe's note book and the CRM system is an opportunity.

Henry: Writing this document as a way of engaging with the database community would make it a very different document. But pointing out the fact that there's a tradeoff that you don't get the integrety constraints that a typical database would provid.e
... But in this case that's a feature, it means you can continue merry along until you need to join the data.

Noah: Another thing we might want to include is some XML.
... It would be possibly confusing to customers if we don't explain how this differs from the story we told about why XML was the answer to this problem.
... We should explain that Acme has invested in XML, and explain why web services and XML don't solve all this problem, because it's not as general a solution, or something.
... It would be helpful to hint why RDF takes it up a level
... In XML the final integration will be custom logic
... If we can't explain this then we have to question why RDF is in the picture.

Henry: In the 2nd or 3rd paragraph, we sort of knock our own technologies by making HTML and XML sound proprietary.
... At the level of elements and attributes, XML is neutral. At the level of triples, RDF is neutral. But once you design a language or an ontology, the underlying technologies become more application specific.
... There's a spectrum between very carefully designed document types that appeal to human beings and purely mechanical representations.

TimBL: The RDF data isn't tree structured so it's always more regular.

Noah: Isn't the same thing in RDF a question of which are managed directly in triples and which are inferences over several triples.
... The question I think is, how did the relational folks sell relational databases?
... Ironically, what came before were hierarchical databases.
... The selling point was relational will probably be slower for a while but it will support ad hoc queries without roto-tilling the physical structure of the data.
... It feels to me that there should be an analog here.
... The selling for RDF has to be that of course XML takes you this far, but RDF takes you this much further. Publicly available data sources, ad hoc query tools, or whatever it is.

TimBL: Suppose we address this in the epilog.
... We can lift the hood and show how sparql queries were done to address the data in the maps.
... and show how it would have been equivalent to some enormous SQL query.

Noah: I think you can do better than that. I would have thought that what's interesting is the fire brings a need to answer all sorts of new questions that they didn't know they had before.
... To the extent that these questions could have been anticipated, other technologies might have been good enough.
... I think that just selling how clean the query is is down in the technology space.

Ed: The problem here is that if you've got three data sources, how do you update a phone number.

DanC: That's the argument not to do RDF. The argument to do RDF is that you don't have to get agreement on the global database schema

Noah: I think of RDF as the projection from the central data source.

<DanC_lap> (anarchic scalability, Fielding calls it. I'd like to do more work on economics of deployment; I sorta hope to use TAG time to do it.)

Some discussion of the data management issues

Noah: Is there research for RDF pub/sub?

<Zakim> DanC_lap, you wanted to note that what's architecturally different between RDF and XML is monotonicity, so that partial understanding is always safe and scalable. Also, RDF

DanC: You can do some of this with XML adapters.
... But when you use RDF adapters, you get partial understanding.
... You can put URIs everywhere in XML, but it doesn't seem cultural.

Noah points out an IBM demo given at XML 2006 in which Atom feeds were automatically generated from XQueries done against XML stored in DB2.

Noah: Atom is becoming the general format for distributing an array of results.

Some discussion of three levels of processing that the scribe didn't really capture

<DanC_lap> (note to self: take another look at Atom vs SPARQL results format, and IBM's SQL/Atom export)

TimBL: Part of the difference, coming back to the parent arc vs. child arc, is that they're both true. Even if half the people do it one way and half do it the other, you can make sense of all of them.

Noah: I would have thought the question is what if someone makes it a child and someone else makes it a grandchild.
... This happens in XML where you need to know the nesting.

TimBL: You can introduce the notion of ancestor and you can say that both child and grandchild are both ancestors and a sufficiently powerful system will understand both.

DanC: The critical point was when you said that you had to know the variability was there and it was safe to operate over it.

Noah describes a two-level graph.

scribe: I don't think I can say "find me any descendant" and know that it'll be correct.

Noah: When I write a SPARQL query, don't I have to know about the ontology?

DanC: Yes, but what you do is virtualize the RDF so that you always know the grandchild link is there.

TimBL: You can combine all the RDF together in ways that you can't combine XML.

Noah: It seems to me that 40-60% of the story could be done with XML

We need to make the merging/monotonicity story clearer.

Noah: What isn't clear to me is, why RDF makes finding the answer faster than finding the answer with XML.

<scribe> ACTION: Ed will try to outline or sketch a story

Vincent: So what are we going to do with the story?

TimBL: Lots of feedback: continue story about dealing with a wider web

DanC: it's written at a different level than the webarch document

Norm: If there's agreement that this goes in a good direction, we should consider telling a larger story, adding some more angle brackets, describing the XML/RDF distinction, and looking at the epilog

Noah: I think we could go farther by having two of the internal data sources using XML.
... Trying to use the same approach using only XML would require certain technologies.
... The good news is they've also used RDF so things are better.

<Noah> So, they are using XML-based integration technologies today to solve some problem, e.g. to combine some data in real time that they put out on their web site.

<Noah> Indeed, it's a good solution for that. There are very right XML-based tools that help them do things like combine the XML and present it as HTML.

<Noah> They might have been tempted to use just XML as the integration technology for all the information we've discussed: from the PDAs, the CRM system, etc. When the fire happened, they would have had to do XXXX; that's much better than what they had before XML, but not nearly what they can do if they wrap all their data, including the XML, in RDF.

<Noah> With the RDF approach, they can do YYYY and ZZZZ that will save their bacon and allow them to be back on the air the morning after the fire.

<Noah> I would like to better understand the comparison off XXXX vs. YYYY+ZZZZ, but I think that's how you motivate that RDF gives compelling advantages over what we have today.

David: I don't see anything in the manufacturing story that's RDF.
... I just replace RDF with XML and it's the same.

Norm: You need to read the punchline.

David: That's an awfully long way down.

DanC: I think you need to put the thesis at the top.

TimBL: We haven't looked at evolution over time yet, but that makes the story even longer.

David: Putting the "aha! lightbulb" further up, that would be useful.

<scribe> ACTION: Tim and Norm to produce a new draft by end of January

Vincent: Tim, was there more on the semantic web that you wanted to talk about?

TimBL: This isn't really a finding, it's just setting the stage

Norm: I'm hoping we fill in technical bits between the stories until we get something webarch sized with stories and technical bits.

DanC: For this meeting, I'd like to look at some of the detailed questions like namespaceDocument-8

TimBL: Yes, I'd like to talk about semantic web versioning.
... maybe we can have a later story

DanC: Yes, stories about versioning and two-hops vs. three

David: So a part three of the versioning finding?

DanC: The three parts I see are: data binding, document like stuff, and RDF
... Document like, hypertext, stuff is a lot of the tag soup issues

David: I don't know how far people want to go with this: taking a look at versioning within particular languages could be useful too

TimBL: Something that does need explaining is that in the SemWeb you get lots of identifiers.

The salient point being that in the SemWeb things get more than one identifier.

DanC: This is "dealing with the cost of aliases"
... because the WebArch document says don't use aliases.

Some discussion of the Microsoft identity stuff and the fact that you want to have multiple identities

Some discussion of the privacy of US social security numbers.

<DanC_lap> namespace document stuff... http://www.w3.org/2004/01/rdxh/spec#mechspec

DanC: Can I talk about some GRDDL stuff that came up as a transition to nsDocument-8?
... We went into one of the mechanic rules earlier.
... a mechanical rule is one of the formulas from the document.

DanC describes the use of identifiers

rdf/fn/grddl are all pretty straightforward

The hard one is dm: which doesn't have a real URI

Norm: The problem with dm: is that there's no namespace name for the dm: prefix in the XPath2/XQuery data model document

TimBL: How hard would it be to do an ontology for the datamodel?

Henry: Given that there are seven, it's probably quite hard.

Noah: 98% of the time people say they have an infoset it's because they parsed or could have parsed a document.
... The infoset doesn't make those constraints, so some of the properties are unbounded.

Norm: The data model is screwed down a little tighter than the Infoset

<ht> http://www.w3.org/TR/xml-infoset-rdfs

Henry: There is an semi-official infoset/RDF mapping

Some discussion of http://www.w3.org/2000/10/swap/infoset/infoset-diagram.svg

<ht> http://www.ltg.ed.ac.uk/~ht/new_xml.owl

<DanC_lap> ^ ht's work of last sep

Henry says this is an ontology for the infoset with cardinalities as of last September.

<ht> That owl ontology is an updating of the one in the Note from Richard Tobin, URI further back above

DanC: The point here is that useful things should have URIs and when they have URIs we will reuse them for crazy things like rules in our GRDDL document.


<timbl> scribenick: timbl

Norm: It was difficult to page this in ... talked with Jonathan and Sean Palmer
... Some has been discussed on ww-tag

<DanC_lap> RDDL: new natures Jonathan Borden (Sunday, 10 December)

Norm: Talked about putting back in the bit about NS documents being documents to be human readable

timbl: I do not feel that that should be overriding

Norm: This is a should, not a must - you can argue that RDF i smore important than human.
... Part 2 was to add transformations to implememty the model we agree on
... Then part 3 as to change the purposes and natures to something more meaningful.
... The curent usage, where a ns doc's URI is used a nature in RDDL has raised eyebrows.
... Jonathan has pushed back and said that the NS name sd thr rddl:Nature is very valuable.


DanC: I see various odd things in RDDL; Are we designing/endorsing this in te TAG?

HT: one position on NSDocument-8 is to endorse RDDL, so this is a TAG matter.

Noah: I am not happy with natures being NS documents -- it is a category error.

DanC: The inconsidtency is the problem: sometiimes it is a namespace document and sometimes it is a class of document.

[whiteboard i scleared ...Norm and VQ have photos of the now erased Acme data integraton diagram]

HT: If things are distinguished from everything else by th namespace, then that is a good thing to be uised a s a naure, it seems.

Norm: However sometimes these things have XML ns names, sometimes the nature is not that of an XMl document at all.

HT: RDDL is intended for use in namespace documents.

Noah: Let's assume that all RDDL documents serve that role: definitional or metadata about the namesapce. Then what question is it answering?

Norm: "What is the nature of the resource [in this namesapce]?"

Noah: So some properties of the namespace will be schema and namespace-like, and others won't
... It should be called the Namespace Description Language really.

[we hunt for the rddl: namespace spec]

@prefix rddl: <http://www.rddl.org/>.

Norm: http://www.rddl.org/nature is 404

DanC: Then I cannot endorse it

Norm: That is just because Jonathan hasn't done his administrative work... I understand there may not be consensus


HT: Let us return to an example.
... There is a document <http://www.w3.org/XML/1998/namespace> which is an HTMl RDDL document in the sense that it has rddl data in it

<rddl:resource xlink:title="XML Schema" xlink:role="http://www.w3.org/2001/XMLSchema" xlink:arcrole="http://www.rddl.org/purposes#schema-validation" xlink:href="http://www.w3.org/2001/xml.xsd">

<div class="resource">

<p> An <a href="http://www.w3.org/TR/xmlschema-0/">XML Schema</a>

<a href="http://www.w3.org/2001/xml.xsd">fragment</a>

is available which constrains the syntax of <code>xml:lang</code>, <code>xml:space</code>,

<code> xml:base</code> and <code>xml:id</code> in the schema language defined by the <a href="http://www.w3.org/TR/xmlschema-1/">XML Schema</a> Recommendation, Second Edition,

of 28 October 2004.</p>




<raman> anyone there?

<Vincent> yes, raman. We are having lunch

<Vincent> It's a working lunch. The topic is namespaceDocument-8

Raman: the current format is that people are running around getting sandwiches and are somewhat clustered around a whiteboard where we are having an informal discussion relating to rddl natures. Not very near the phone. I doubt you'd hear well, but we can dial in if you like. Suggest we do that when the real meeting resumes, which is likely to be in 10-20 mins when Stuart gets off the road. I think we're going to talk to him about F2F scheduling.

[DanC and Norm discuss DocBook as a specific RDDL example. ]

Lunch revealed an consensus among danc noah and timbl that hte current rddl usage of sometimes a DRENS (doc root elemnt namespace ) as a rddl:nature and sometimes somthing quite other as a category error which seiously limit sthe use of the metsdata.

TimBL proposes in the RDF model for RDDL data the rddl:drens as a property of the (schjema, for examle) documents inked froma RDDL document. It can be fished out and verified automatically, and provides the semantics of the type (nature) of a resource... well defined semantics. other non-xml documents such as ISO PDFs could have classes , maybe with names made up by RDDL, or could have language-spec-publisher-homerpage links, etc.

we reconvene

HT explains that his code uses both purpose and nature to find a resource to use in validating a document, since Schema decided to support RDDL.

[see photo of board]

<vq> whiteboard, see also photo by Norm

DanC seems to be proposing a situation in which a language is identified indirectly by giving the URI of the gammar which defines it..

<..../ns/docbook> normativeSpec <../docbook.html>; spec </docbook>; is language of L; namespaceTransformation <grddl.xslt>; htmlRenderTranform html.xslt; FORenderTransform <do.xsltr>;

informativeSpec <../docbook.pdf>.

<>../rnc> content-type mt:rnc # @@


<..../ns/docbook> namespaeTransformation <grddl.xp>. <grddl.xp> contentType mt:XProc.

<raman> dialing ...

<raman> you guys are not dialed in.

Norm: Looking at that model, if I want an XQ transformation which applies to a document -- I have to invent or standardize a new URI for XQuery.

Danc: Yes, as you do with RDDL at the moment

Norm: Well, I ahve thc onvention of using the DRENS.

DanC: I am happy for you to give a URI to that to formalize it.
... The properties in the diagram are very crisp, well defined meanings.

logger, pointer?

HT: My way of expressing what I think is norm's concer:
... We are dsigning this ontology against an unknown set of requirements.
... You are chosing this structure based on good pracyice and exsiting experience, but not based on defined RDDL requiremen ts, and I think it makes these
... I can;t read off the things from this diagrm which i can from the RDDL document.l

[discuss case of 2 xslt xforms one of which makes html and one of which makes FO]

Now the case of XSLT as a schematron validator ... the content-type is XSLT, same as the content-type for an XSLT presentation transform to HTML.

HT: Microsoft have cde which consumes this.

Norm: I would like the RDF model to be nice and simple, which RDDL users could understand.

TimBL: I like drens as a concept to add in, as it can be genarted automatically, is easy to allocayte, is useful info.

HT: Can we ansswer --- if it were only the Xlink:role whcih is schizophrenic, and not the RDF , and if there was some detreministic algorithm for constructing the RDF form the attribute, then we could have simpl ontology

Norm: Docbook 4 has no namespace, Docbook 5 has a namesapce.

<raman> I'm about to hang up because this is completely impossible to partipate/follow

Looking for RDDL examples

HT: Two :

- the ns dco for XML itself

- large quantities of documents inside MSoft according to Paul Cotton.

- - and this doument:

<ht> http://www.w3.org/XML/Group/2005/12/XMLSchema.html

Norm: I was going to write a GRDDL transform for RDDL whcin it looks simple.

HT: The purpose of using these things is you look fro strings you know

Noah: Still, namespace name is not the nature of the document: that is just a category error.
... Ontologically the things are quite different.

Norm: Examples of situatins which have > 1 schema, both validating, but not the same level/typ of validation. boht in rtelaxNg

ACTIION Norm to provide a set of test cases of whays in which RDDL is actually used

DanC: DocumenElementNameSpace

[discussion of a URI for that]

Noah: documentElementNameSpace


HT: documentElementNamespace

Tim: put it in an XML namespace


<DanC_lap> ACTION: Norm to provide a set of test cases of whays in which RDDL is actually used

<DanC_lap> ACTION: DanC to start an ontology including docns/documentElementNamespace

TAG face-to-face meeting schedule (cont)

VQ: looking for TAg face-face meeting, we had March 1-2 but Stuart can't come then. Location ws MIT

Norm: Tim couldn't travel more around ten, but also there was a W3C workshop the 3 days before

HT: Propose move it to Monday 5 and Tuesday 6 March

6 and 7 to keep clear of weekedn ?

<Stuart> I *can* do 6-7 March, Cambridge MA

<dorchard> yuck to 2 trips to boston then a trip the following week.

<Norm> do you have an alternative to propose, dorchard ?

HT: We have nothing better than 6-7 March Boston

<dorchard> I could do UK the week of Mar 12

<dorchard> I could host in vancouver week of 6-7 Mar :-)

<DanC_lap> (how is 6-7 March in BOS better than 1-2 March in BOS?)

(Staurt can make it)


<DanC_lap> (thanks)

<dorchard> ( I think stuart can make that)

<raman> cant do UK

<DanC_lap> SKW: I'm unavailable 28 Feb +/- 1 day

<ht> Stuart cannot do 28 Feb or 1 day either side

<Noah> Noah and Tim can't do week of Feb 19-23

<raman> we've made no progress?

<DanC_lap> PROPOSED: to meet 5-6 Feb at MIT, contingent on TimBL's availability

<DanC_lap> PROPOSED: to meet 5-6 Feb at MIT, contingent on TimBL's availability, noting 5-6 March fallback

<DanC_lap> PROPOSED: to meet 5-6 Feb at MIT, contingent on TimBL's availability, noting 6-7 March fallback

<DanC_lap> TVR: I'm not available for 5-6 Feb

DaveO can't do March 6-7

<DanC_lap> TVR might be available 8-9 Feb by video

<DanC_lap> PROPOSED: to meet 5-6 Feb at MIT, contingent on TimBL's availability, noting 6-7 March fallback. VQ to contact the candidates. noting TVR's regrets.

<raman> parallel audio channels: Vincent/henry/Noah ...

<DanC_lap> ACTION: VQ to collect availability information from candidates about TAG meetings (1) 5-6 Feb at MIT or (2) 6-7 March at MIT

Issue RDFinXHTML-35

Discussion o action items ... DanC's re GRDDL

DanC: There is a companion to the GRDDL spec, which could be a TAG finding.

<DanC_lap> Storing Data in Documents: The Design History and Rationale for GRDDL

RESOLUTION: Put Issue RDFinXHTML-35 into a Pending state pending work by the GRDDL WG and/or the RDFa/HTML/SemWeb-deployment WGs

<DanC_lap> pending aka exit:deferred

HT: I think that the CDF group is NOT working on what we expected them to work on, but is doing the by-reference version of compiund documents, which is NOT the first think one thinks of when it comes tro compound document architectures.

Raman; I think that this is important to one of the vendors in the group.

Issue xml11Names-46

HT: The XCHema, etc etc have addresses the concerns about XMNL 1.1 new unicode characters, and so this concern has been addressed effectively

Norm: I handed some ball to the XML CG ... can't remember exactly what that ball was.

<DanC_lap> DanC: how about a testcase?

HT: I do 't remember it in HTML CG recently

<DanC_lap> HT: yeah... the icecream case...

<Norm> ACTION: Norm to find a test case; reminds self to check with Henry for the Ice Cream example

HT: The example i remmeber is the dutch IJ character which is used in the words for Ice Cream

DanC: Can we have a test case, maybe?

Norm: Propose we drop my action if I take a new one to make a tets case

<DanC_lap> ok by me

ok by me

<ht> The Dutch word for icecream is "ijs"

<scribe> ACTION: NW to accepted on 4 Oct 2005: Check on current status of issue xml11Names-46 with XML CG [WITHDRAWN]

<ht> http://www.w3.org/TR/2005/NOTE-xml11schema10-20050511/#d0e138

<ht> http://www.w3.org/TR/2006/WD-xmlschema11-2-20060217/

<DanC_lap> [[

<DanC_lap> XML Schema 1.1 Part 2: Datatypes

<DanC_lap> 17 February 2006, C. M. Sperberg-McQueen, Ashok Malhotra, Paul V. Biron, David Peterson - (Royalty-Free Commitments)

<DanC_lap> Last Call Ends 31 March 2006

<DanC_lap> ]]


<Norm> <ijce_cream/>

<DanC_lap> (did HT say "we don't do test cases" on behalf of the XML Schema WG? ouch.)

<DanC_lap> (ht clarifies to "we have not".)

<DanC_lap> (as in... yet.)

<DanC_lap> ACTION: HT to accepted on 7 Feb 2006: Bring us back to xml11Names-46 after the XML Schema WG publishes its expected Last Call WD. [DONE]


Concluding Remarks

Round the table about what ine is excited aboit ebing about top contribute

DanC: versioning, tagsoup, GRDDL (in that group). I mentioned some news on siteData, but don't have much energy to push it

TimBL: Sem Web story ... adding more story and maybe some diagrams like the one on the board and s reen shots ....
... Versioning happy iwth the group's progress there

DaveO: Versioning between now and our first face-face .... depending on election .... URNs abnd Registries and teh State finding both on the back burner alas.
... i think TAG oup i very important to the TAG.
... Web sevices and wbe integration is important to. Think we should do some web 2.0 sgtuff .... dpnm't see ho to get the TAG started.
... The AJAX, uFormat, etc, which should be related to Web Architecture

Raman: I would like to make aliot of ht the W3C and TAG do more relevant to what we do on the Web.
... I don't think we have answered aquestion [missed]

<DanC_lap> (hmm... on innovative stuff, my mind goes toward less let's-write-one-document-together to individually blogging articles. any interest in PlanetTAG?)

Norm: Interetsed in flushinng out the Sem Web stuff, and looking more that the Versioning issue.
... I feel deeply connected to the TagSoup issue. Also, the whole identity on the web thing
... i would liket to push forwatrd,

HT: I have => 3 things pending.
... 1) URNs adn Registries we need to finish and get iout there
... 2) The formal semantics of XML documents as discussed in Vancouver: my line on Versioning Part I, combined with my and DanC and Pat Hayes' intuitions on that
... 3) We need to brainstorm architectural level innovation for the TagSoup issuee. We are well-placed to do that there. I think we have a lot of talent and no built-in bias... I would truste each TAG member (and candidate) to talk dispassinately about this issue .... many people are very entrenched, but not TAG members.

<DanC_lap> (I heard HT say we were NOT close to having a finding there.)

Vincent: I was interested in eth sem web architecture based on the new story we dscussed today. I think we realy need this comprehensive view of how we use these technologies to fit tghetehr and solve real issues.
... i am slao interested in the TagSoup nightmare.
... If we could have a few directions for addressing this complex issue.

Noah: ... I have prrtty much wrapped 2 finings, metadtaInURI and rule of LeastPower.
... (Larry noted we misquoted the history of PDF and PS but we must fix that )
... *(If I can fix my cygwin then I will be much more effective)
... In the net few months I would propose to spend my time split.
... 1) A new major theme for my investments ... ie DO has had Versiuoning. wonder about schemeProtocols.
... Maybe I have learned enough to contribute
... 2) I would like to contribute to Dave's work on versioning.
... 3) Enterprise workshop stuff.
... joined the TAG at a time when there was discomfort about TAG direction. The TAG had just released AWWW and didn't know how to continue
... Now we are doing findings which is fine.
... I find though that the TagSoup issue is as much social as technical issue. I think we can re-establish credibility by contributing to that . No one is waiting for our answer though.


<Zakim> DanC_lap, you wanted to note tlr's interest in the state finding

DanC: [refelcts] Thomas Roesler has shown interest related to State fidning
... Also the Princple of LEast Power finding was a really useful referent in a discission eg in Austin



Summary of Action Items

[NEW] ACTION: Ed will try to outline or sketch a story
[NEW] ACTION: Tim and Norm to produce a new draft by end of January
[NEW] ACTION: DanC to start an ontology including docns/documentElementNamespace
[NEW] ACTION: Norm to find a test case; reminds self to check with Henry for the Ice Cream example
[NEW] ACTION: Norm to provide a set of test cases of whays in which RDDL is actually used
[NEW] ACTION: VQ to collect availability information from candidates about TAG meetings (1) 5-6 Feb at MIT or (2) 6-7 March at MIT
[DONE] ACTION: HT to accepted on 7 Feb 2006: Bring us back to xml11Names-46 after the XML Schema WG publishes its expected Last Call WD.
[DROPPED] ACTION: NW to accepted on 4 Oct 2005: Check on current status of issue xml11Names-46 with XML CG
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2006/12/22 10:49:58 $