---------- Forwarded message ----------
From:
Drummond Reed <drummond.reed@xdi.org>
Date: Thu, Dec 9, 2010 at 8:04 PM
Subject: Minutes: XDI TC Telecon Thursday 1-2:00PM PT 2010-12-09
To: OASIS - XDI TC <
xdi@lists.oasis-open.org>
Following are the minutes of the unofficial telecon of the XDI TC at:
Date: Thursday, 09 December 2010 USA
Time: 1:00PM - 2:30PM Pacific Time (21:00-22:30 UTC)
ATTENDING
Giovanni Bartolomeo
Markus Sabadello
Drummond Reed
Joseph Boyle
Michael Schwartz
THE GOTOMEETING FOR TODAY IS:
https://www2.gotomeeting.com/join/969244355
THE IDEARPAD LINK FOR TODAY IS:
http://xdi.idearpad.org/15Please try to preface each of your comments with your name so the
transcription into the minutes is easier.
1) REVIEW OF XDI GRAPH PATTERNS
The entire topic for the call was the patterns that have been posted at:
http://wiki.oasis-open.org/xdi/XdiGraphPatternsIn particular, Giovanni has posted several alternative patterns, and Drummond has posted the proposed link contract pattern he has been working on.
We began by walking through the page reviewing Giovanni's comments.
His first comment was about using self-referential arcs to indicate comments. The self referential arc for Giovanni violates the mereological principle he has followed to assign semantics to the XDI graph [discussed in ref-minutes]
Drummond notes that there is no disagreement about what a synonym is, and Giovanni's suggestion is intended to make it clear there's no semantic difference.
# ACTION: Drummond to suggest a way to express synonyms for a root context node.
Giovanni's second comment was about the assertion of authority for an XDI document.
Drummond replied that he has been thinking about a stronger mechanism. He believes that they are two separate issues: 1) how to express synonyms for the root node, and 2) how to assert authority for an XDI document. He believes the latter is better handled by a link contract.
We talked about how XDI discovery works. Drummond said his view of XDI discovery is that there is one logical global XDI graph, and then many physical instances of that graph, and that XDI discovery is starting with some known instance and making XDI $get requests from that graph to discover the URI of the instance that has the target portion of the graph you are trying to reach -- and iterating this process as needed, just like in DNS.
Giovanni said that he believes that there is a similarity, from a logical PoV, between discovering those instances in the global graph and discovering, for example, a subgraph in a reification sentence (e.g. discover subject =bob in the assertion =alice/+say/(=bob/+age/"31") )
# ACTION: Drummond to add illustration/explanation of how he sees XDI discovery working.
We discussed the use of the word "fractal" with regard to the XDI graph. We agreed that each instance of the graph is a copy of the same structure of the whole -- and it that sense it fits -- but that each instance is not a copy of the content of the whole -- and in that sense it doesn't fit, because in that sense it is not self-similar.
********
The next discussion was about the Single-Valued Property Pattern. Giovanni proposed a different pattern, labelled on the wiki page as "proposal".
He walked us through his rationale for the proposed pattern, which is also what came out of the Whistler retreat. It contextualizes the property +age in =abc, subjectifies it and declares one of its properties ($d) in a similar way as you assign a value to property +age of subject =abc.
Drummond pointed out the core semantic problem with Giovanni's proposal. To assert a datestamp on a +age property, you must define "a property on a property". That is what we are trying to model in the Single-Valued Property Pattern diagrams. In order to model a property on a property, the "subject" of that metadata statement must be a property. In the diagram on the left, the "subject" of the $d extension is =abc/+age. =abc/+age is a property of =abc. Whereas =abc+age is not a property, but another XDI subject.
Giovanni pointed out that it is not necessary to be a property. In RDF, you can use a property as a subject when you have to state something on the property itsef (examples include rdfs:domain and rdfs:range).
Moreover, Giovanni said that the left graph is asserting something slightly different than the right graph.He tried to figure out another example. Property --> +color, property on a property --> +hue, i.e. +color+hue
=mycat/+color+hue/+grey
=mycat+color/+hue/+grey
+color+hue exists even if =mycat is not existing (you could even have +color+hue as subject in another statement, assumed there is something to say about +color+hue). You can apply then the first assertion (asserted in the left graph); in the second statement, you suppose there exists a cat (=mycat), and it has a color; then you state some property on =mycat's color. Open issue: could we have similar thoughts when we think about +age$d?
^^^^^^^^^ "left" AND "right" WERE ORIGINALLY REVERSED IN THE PARAGRAPH BELOW. THE VERSION BELOW IS NOW CORRECTED ^^^^^^^^^
Giovanni pointed out that it one key semantic difference is the scope of the assertion about the property. In both diagrams, the scope of the property +age is =abc. On the right diagram the scope of the property $d of the property +age is strictly within the context of =abc. On the left diagram, the scope of property $d is strictly in the context of =abc+age, i.e. exactly =abc's +age (and not e.g. =xyz+age).
Joseph brought up that if there is going to be any semantics involved with trailing slashes, they should be modelled as nodes.
2) NEXT CALL
We agreed that we would hold calls both next Thursday (although Drummond may possibly have a conflict) and Thursday the 23rd, and probably not hold one on Thursday the 30th (to give us a week off between Christmas and New Years).
------------
ONGOING ISSUES LIST
Each of these is a candidate for the agenda for future calls.
* PROPOSED CONSTRUCTS/OPERATORS FOR XDI
Discuss the following wiki page originally posted by Giovanni:
http://wiki.oasis-open.org/xdi/XdiNewFoundation* DICTIONARY STRUCTURE
Mike would like an example of the PDX dictionary as soon as we can do it.
* EQUIVALENCE SEMANTICS
Close on whether we need an additional $ word that is the equivalent
of Higgins Personal Data Model (PDM) semantics of h:correlation,
which is not as strong as $is.
http://lists.oasis-open.org/archives/xdi/201006/msg00036.html
* COOL URIS
Continue previous discussion about the use of standard RDF URIs in XDI:
http://lists.oasis-open.org/archives/xdi/201006/msg00023.html