<scribe> scribe: Scott

Meeting Minutes from April 11

<tbaker> +1 accept minutes

Proposed: accept minutes

Resolution: accept the minutes from last night.

action items

Guus: Gavin do you want to record a new issue.

Gavin: I'd like to but can't

Turtle LC

Guus: Gavin can you comment

Gavin: I should be able to get the remainder of the minutes in my the next conference

Guus: Just editorial comments no major new issues.

<gavinc> application/n-triples vs. application/ntriples

Gavin: Do we want a dash or comma in the media types?

Guus: no dash you mean. that's 12.1

Gavin: The dashes are used like segment markers.

… not able to find languages that had dashes

… we don't believe any application uses slash n-triples at the moment.

Guus: My intuition -- leave out the dash.
... Sandro can you comment.

<MacTed> http://www.iana.org/assignments/media-types/index.html

<gavinc> example with dash: audio/x-ms-wma application/x-gzip video/x-ms-wmv

Sandro: Maybe a dash is used like a space.

<gavinc> application/x-font-ttf

<sandro> http://www.iana.org/assignments/media-types/application/index.html

Gavin: there is no code that uses an n-triples media type.

Guus: what will be easiest to remember.

Sandro: how do we write it in the spec.

Gavin: We write it with a dash.

Sandro: we call it n hyphen triple

<sandro> cwm -n-triples

Gavin: others use it without a dash. use it on the command line.

<sandro> cwm --n-triples

<gavinc> cwm -ntriples

<PatH> Sorry Im late.

Sandro: Should use two dashes if more than one char.

<davidwood> +1

<gavinc> +0

<Souri> +1

<sandro> +1 dash (mildly)

<MacTed> +1

Guus: type a plus one you are in favor of the dash.

<ericP> +1 to —

<gavinc> heh

<sandro> +1 to &dash;

<gavinc> Entity 'dash' not defined ;)

<sandro> love some of these mime types, like: vnd.openxmlformats-officedocument.drawingml.diagramLayout+xml

<MacTed> ietf-types@iana.org :-)

<gavinc> RESOLVED: Use application/n-triples for content type of N-Triples

<ericP> ACTION: ericP to send mail to ietf-types to request the media type application/n-triples [recorded in http://www.w3.org/2012/04/18-rdf-wg-minutes.html#action01]

<trackbot> Created ACTION-163 - Send mail to ietf-types to request the media type application/n-triples [on Eric Prud'hommeaux - due 2012-04-25].

Guus: Part one has been covered before.

… plan for a new draft next week.

Gavin: another week would be nice.

<MacTed> ( mimetype hyphen appears likely acceptable -- http://www.iana.org/assignments/media-types/application/remote-printing )

Guus: We recorded a new date 2 may for the new draft to be available

Named graph semantics

<PatH> whre is that new use case mentioned?

Guus: Ask to write UC Europeana to use case page.

Sandro: I want to understand use cases as a coder. Discussion on this?

Guus: I'll ask one of the authors to join the telecom.

Sandro: ask him to write it first.

… in something a coder can understand.

Guus: It's an enormous amount of data.

… maybe you can ask him a few questions on the list.

Sandro: ok

Guus: They have issues. Art work described by multiple organizations.

subtopic: RDF with Contexts

<sandro> http://www.w3.org/2011/rdf-wg/wiki/RDFwithContexts

Guus: Thanks to Pat for the wiki page.

… We should start with QA

… session

<davidwood> +1 to PatH for writing this up

Pat: the generalization if of something not necessarily a graph. Deliberately don't restrict the form of that document.

<gavinc> +1000 allowing it to be empty

<davidwood> +1 to allowing empties

… agree on the meaning of the IRIs without specifying what that meaning is.

Sandro: when will two colleagues publishing about the same thing use IRI's in different ways.

Pat: Data from different sources should be in diff contexts ( some have this perspective)

Sandro: There is a global hypotheses,

Pat: RDF now accepts the global hypothesis.

<davidwood> Accepting a local hypothesis makes *no statement* about the global hypothesis.

Sandro: If the global hypothesis holds you don't need this.

Pat: Still very useful to put one context under another.

… recording this context more can be added.

… one URI can mean progressively finer things as people add to it.

… even in the global hypothesis things can work this way.

Sandro: an example?

Pat: You might rely on a more restrictive meaning accidentally.

… easier to use the same term with some restriction.

<Zakim> MacTed, you wanted to say the global hypothesis has been disproven already... the challenge is dealing with data that is presented as if it were valid

Ted: the global hypothesis is the ideal but most are not living by it. Local is the way things are.

… it has been disproven. Giving up is appropriate. Lots of data is produced based on it. we have to deal with it.

<ericP> i'm not sure why I use terms like foaf:name if they mean different things to different people

<ericP> what's the point of RDF anyways?

Sandro: The same URI is supposed to mean something different for the two us?

Ted: We may both be wrong. This is the case with owl:sameas.

<sandro> q>

<Zakim> davidwood, you wanted to provide other examples of violation of the global hypothesis

<pfps> how is owl:sameAs broken???

david: I agree with Ted:

<sandro> sandro: The way to get the genie back in the bottle is Running Code. We'll get there some day.

<Guus> I think the poinnt ewas not that owl:sameAs is broken, just the way it is used

… Data from the data web shoved into a triple store. Using these I use some and pretend others don't exist.

… you're overriding data that you choose.

Sandro: Give an example:

David: Import from DBpedia. RDF label in a dozen langs. I strip some or all and replace while reusing the rest of the triples.

Sandro: Global, when you see rdf: label you know what it meant.
... Would the new label to be true?

David: I think so. I'm gong to make it up or do what I want.

Sandro: That doesn't mean you disagree with the use of the word.

Pat: I think you are right but how do you tell the difference.

Sandro: You can handle all in the first possiblity.

<davidwood> What do you mean by "God"?

Pat: You are presenting an extreme local view.

Sandro: how can there be a middle ground?

<Zakim> ericP, you wanted to ask how we bridge from conventional RDF to a local hypothesis RDF

<sandro> sandro: Any time you think you belong on the same island, then you could both use the same URI in the global-hypothesis world.

Eric: What mechanism will allow me to merge rdf graphs.

… how would localities agree on how subjects or predicates mean the same thing.

… once the locality is introduced does it require more rigor to make the query.

Pat: A context must be agreed on for a given group.

Sandro: If a larger group -- do they add another context URI.

Ted: There's a lot of data out there without context.

… we'll have to deal with the fact that it is produced in one context and used in another.

… The assumption is a dirty merge.

… global does not work because we don't have a clean, solid reference of what that is.

<davidwood> ..and the dirty merge is a feature, not a bug. It is critical to the flexibility of RDF.

Pat: If people pretend it does. It will go on as before.

… choice can be made to put them into contexts. Just put inheritance tags in place.

<Zakim> sandro, you wanted to say people can be wrong, people can be talking about different points in time, people can be talking about imagined universes, yes. But we can't just give

Sandro: Anybody who wants to merge with someone else's data has to have a special agreement.

Pat: Tell the world that they have to be careful, so they can check to see if their data is consistant.

Sandro: Vocabularies can allow mix of 14 different properties.

Pat: a Large ontology has used this successfully. You don't get exponential explosions. They resolved ambiguities by introducing contexts.

… they introduced a relationship of covering.

… these things have something in common as defined by a top context.

Sandro: Theres a super property coving the different senses.

Pat: A change in context shifts vocabulary meaning.

S

Sandro: that makes a lot of sense from an NLP perspective.

<pfps> I'm still having a hard time figuring out just why contexts should become part of RDF.

… concerned about what it means in publishing RDF graphs, because you don't know the vocabulary.

Pat: It should be used in a context that changes meaning intended.

Sandro: a little can be done in existing OWL.

<Guus> q?

… a person has a bunch of properties. A person is a professor, there is new meaning. OWL reasoning understands that.

Pat: You can do things like age changes.

Sandro: I'm concerned about vocabularies changing meanings arbitrarily.

<ericP> +1 to "Hieronymus Bosch" description

Ted: Meanings are going to shift regardless.
... AAA means this is going to happen.

Eric: Introducing of islands of data that I don't trust. No force will stop people from stopping RDF used as now.

… semantics entails new use cans but the complexity doesn't motivate use.

… I don't seem context as being marketable.

Pat: It's just one new RDF relation.

Sandro: Everyone in a given group could give a context.

Eric: Adding contexts gets too complicated.

<Guus> Pat: more than 1 context probably means union (not intersection)

… many people want to share terms, there is a transitive connection. Two predicates written by different people and a third connects them.

Ted: A wrong word meaning can be wrapped. Code can swap it out.

Sandro: this doesn't solve change over time by itself.

Pat: Change over time can be described as an updating of a concept. The old terminology continues in use but is updated.

… all you have to do is invent a context and insert old stuff.

Sandro: How is this different from changing the URI to point to new data.

Eric: If the graph is sealed, a snapshot is in order.

… we

Pat: No one is saying you have to stop what you are doing.

… mends some things that are currently broken.

Sandro: If you want to say that two graphs can be merged using this context. It means it can't be done otherwise.

… seems like it can't be otherwise.

Pat: the current RDF semantics are the default.

… if you assert in the context, you are being careful. Otherwise URI's are just URI's.

Eric: What it does is it takes the private, illegal formats are being brought into a set of rules where they are.

<sandro> sandro: IF it's stated clearly enough that rdf:inherits is a way to OPT OUT of the global hypothesis, then I think this is an okay experment, and harmless.

Eric: legal.

Pat: It lets inferior uses into the public.

<sandro> sandro: So this is a sort of Amnesty proposal.

Sandro: I'm ok with this as long as RDF inheritance as a way to opt out is clear.

<sandro> sandro: So this doesn't really solve the owl:sameAs problem.....

… (correction) the default is a way to opt out of the RDF inheritance.

David: what should the default be.

Sandro: the namespace document needs to opt in.

Eric: updating the namespace documents is onerous.

David: there will still be unintentional differences.

<sandro> sandro: (switching sides) There's a strong case to be made that people who really mean their RDF, and really read the ontology documentaiton, might opt into that.

<davidwood> I don't think *anyone* doesn't mean their RDF.

<MacTed> and then... the ontology on which they based their RDF gets redefined.

<MacTed> uh oh!

<davidwood> Some people just make bad RDF - by my (or others) definition.

This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/data/date/
Succeeded: s/example:/example?/
Succeeded: s/owl:sambas/owl:sameas/
No ScribeNick specified.  Guessing ScribeNick: ScottB
Found Scribe: Scott

WARNING: No "Present: ... " found!
Possibly Present: AlexHall Eric Gavin Guus MacTed NickH OpenLink_Software P21 P26 P27 P31 Pat PatH Proposed Sandro ScottB Souri Ted Tom_Baker Tony aaaa aabb danbri david davidwood ericP gavinc global hypothesis manu manu1 mischat pfps subtopic tbaker trackbot up was yvesr_ zwu2
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting


WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 18 Apr 2012
Guessing minutes URL: http://www.w3.org/2012/04/18-rdf-wg-minutes.html
People with action items: ericp

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.