12 May 2009

See also: IRC log


+1.781.643.aaaa, mhausenblas, jar, +1.512.342.aabb, David, Alan, TimBL




<mhausenblas> hi there

<mhausenblas> not sure if we have a fixed agenda (didn't see one) ...

<mhausenblas> so FYI, after last telecon I took TimBL's http://www.w3.org/2006/gen/ont# ontology and turned it into a diagrams

<mhausenblas> http://esw.w3.org/topic/AwwswGenOntDiagrams

<mhausenblas> this may help us today


<mhausenblas> http://esw.w3.org/topic/AwwswGenOntDiagrams

<mhausenblas> http://www.w3.org/2006/gen/ont#

What is an example of an IR that is not time generic?

Are time generic and time specific mutually exclusive?

TimBL: Let's try where they're disjoint.

Michael: How about if we look for an example of each of the classes?

<timbl> http://lists.w3.org/Archives/Public/semantic-web/2009May/0071.html is time-specific

<timbl> http://www.w3.org/TR/ is time-generic

time-specific = has same representations at all times?

<dbooth> forall time t1, t2 and any other request inputs i1, i2, f(t1, i1)=f(t2, i2).

A resource can be updated, but if it's time specific, then the update doesn't lead to changes in the representations?


<dbooth> hmm, the rule that I just gave also fixes it over other inputs. Not sure that was the intent.

<dbooth> Okay, lemme try again: forall time t1, t2 and for any input i, f(t1, i) = f(t2, i).

<mhausenblas> wondering if the TAG finding http://www.w3.org/2001/tag/doc/alternatives-discovery.html is of any help (and/or has been taken into account so far)

if time specific, then when i do an http operation, then i don't need to redo a GET, because it's not changing

(that was me recording TimBL)

<timbl> It does highliht the ,medium as a dimension.

<timbl> (print, laptop, mobile, screen etc)

<timbl> mh, the arrows in the diagram at the bottom of the finding above are links, not relationships.

<timbl> into design issues

<alanr> watch out: two spellings (capitalization) of [tT]imeSpecificResource and [cC]ontentTypeSpecificResource

David, your "for all time..." needs to be qualified: the equation need only hold when both are defined.

<timbl> yes, property and class

TimBL: a GR is *not* determined solely by its representations

<dbooth> jar, when both what are defined?

<alanr> http://esw.w3.org/topic/AwwswGenOntDiagrams?action=AttachFile&do=view&target=genont-classtree.png

<alanr> just the class/subclass relations

both f(t1, i) and f(t2, i)

<dbooth> hmm, interesting point.

can the set of all r such that r is a representation of g (for some g) be infinite?

TimBL: Can be.
... When we measure it via HTTP, we can only make a finite number of measurements.

<timbl> E.g my blog can contain an ad which has a random integer generated at run time which is unbounded.

<timbl> One deosn't normally

There is a relationship between the server and the generic resource

<dbooth> jar; What do I learn about the resource by doing a GET?

<dbooth> dbooth: You learn what one of its representations is.

What do I learn about a generic resource by doing a GET?

I learn that the returned entity is a rep of the resource, right?

<timbl> yes

Trust is orthogonal to authority, right?

<timbl> if the returns status was 200

TimBL: Happy to not talk about authority issues just now
... You seem to think the ontology as if it exists in isolation

could someone scribe?

<dbooth> timbl: by sticking RDF in the metadata about a document (indicating that the resource is time-specific IR) then I know that there is no need to fetch it again later.

<dbooth> (paraphrased)

The boundary around an ontology is arbitrary

(that was TimBL)

Sorry dbooth. you were on the q

TimBL: The ontologies aren't in a stack

<dbooth> Here's a little more I wrote for IRC that pertains to earlier discussion:

<dbooth> I have treated ftrr:IR as a function of two inputs (Time x Request), but if desired we could break the Request input into separate inputs such as: Language, ContentType, Client, OtherRequestInput.

<dbooth> Then each *-specific resource would just be holding the representations constant across different values of that input. So a language-specific resource would be:

<dbooth> forall Language l1, l1, and any Time t and ContentType ct and Client c and OtherRequestInput o: f(t, l1, ct, c, o) = f(t, l2, ct, c, o).


jar: Does there exist a GR g such that there is no R with R a representation of g (at any time)


(In Roy's model, the answer is yes.)

TimBL: Boundary case we don't care about.

E.g. something you can PUT or POST to, but you never get a 200

michael: Not sure that's a border case

I get confused because we go back and forth between ontology and protocol

TimBL: It would be mathematically useful to have a GR that no reps.

e.g. no one has written one yet.

<timbl> Not practically useful

Could you define G = the top resource? (every representation is a rep of G?)

TimBL: Yes, but not useful

So utility is not a requirement of a GR.

TimBL: Bad ideas are not outlawed by the ontology
... but in webarch, we say that consistency is a good idea [?]

can't we let the market decide?

TimBL: society will hold you to what you serve according to what it demands

<timbl> ~

<mhausenblas> Michael will take care of organising the next telecon on 26 May

<alanr> Here is a class tree with labels that describe how I understand what the intent is: http://esw.w3.org/topic/AwwswGenOntDiagrams?action=AttachFile&do=view&target=genont-different-labels.png

<mhausenblas> [adjourned]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009/05/12 14:07:21 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

No ScribeNick specified.  Guessing ScribeNick: jar
Inferring Scribes: jar

WARNING: No "Topic:" lines found.

Default Present: +1.781.643.aaaa, mhausenblas, jar, +1.512.342.aabb, David, Alan, TimBL
Present: +1.781.643.aaaa mhausenblas jar +1.512.342.aabb David Alan TimBL

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: 12 May 2009
Guessing minutes URL: http://www.w3.org/2009/05/12-awwsw-minutes.html
People with action items: 

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

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report

[End of scribe.perl diagnostic output]