Minutes: XDI TC Telecon Thursday 1-2PM PT 2010-01-21

[Drummond Reed <drummond.reed@xdi.org>]

Following are the minutes of the unofficial telecon of the XDI TC at:

Date:  Thursday, 21 January 2010 USA
Time:  1:00PM - 2:00PM Pacific Time (21:00-22:00 UTC)

ATTENDING

 

Giovanni Bartolomeo

Drummond Reed

 

REGRETS

 

Bill Barnhill

John Bradley

Markus Sabadello

1) $HAS SEMANTICS

  http://lists.oasis-open.org/archives/xdi/201001/msg00011.html
  http://lists.oasis-open.org/archives/xdi/201001/msg00012.html

 

With light attendance due to travel & conflicts, Giovanni and Drummond decided to concentrate on the discussion about $has semantics. Following is a transcript of the chat conversation that accompanied their telecon discussion, along with annotations from Drummond to help add what was discussed on the phone but does not appear in the transcript.

 

The chat began with a quote from Giovanni’s latest email on the issue (http://lists.oasis-open.org/archives/xdi/201001/msg00053.html).

 

In essence Giovanni said he agrees that:

 

+a/$has/+b  <=>  +a+b

 

but, as Bill pointed out:

 

saying "Bill's email" is somehow different that saying "Bill has an email". Therefore I think that the following are not aliases (synonyms):

 

+a/$has/+b != +a+b

 

Giovanni then gave a few examples of why he believed this was true.

 

[1:27:14 PM] GB: =bill+email/$has/+address

[1:27:39 PM] GB: =bill+email/$has/+size

[1:28:08 PM] GB: (=bill/$has/+email)/$is$a/+statement

 

Drummond then responded with his key question about Giovanni’s proposal that +a/$has/+b != +a+b

 

[1:29:12 PM] DR: so, as I've been thinking about this, here's the question I have: does =bill+email and =bill/$has/+email address the same point in the graph?

 

Drummond then framed one way to think about this, which is in terms of what Bill calls “resolution equivalence”. To discuss this, Drummond and Giovanni used the notation proposed by Bill (recorded by Drummond in the wiki spec at http://wiki.oasis-open.org/xdi/XdiOne/RdfGraphModel#Notation) with the following three XDI statements:

 

            #1: =bill/$has/+email   #2: (=bill/+email)   #3: =bill+email

 

Giovanni and Drummond agreed on the following:

 

            =bill/$has/+email  <=>  #2: (=bill/+email)  <=>  #3: =bill+email

 

Drummond then wondered if Giovanni was saying the following:

 

            =bill/$has/+email  !=  #2: (=bill/+email)  ==  #3: =bill+email

 

Giovanni clarified that no, he is saying the following:

 

            =bill/$has/+email  ==  #2: (=bill/+email)  !=  #3: =bill+email

 

Drummond then asked:

 

[1:40:13 PM] DR: let me phrase this question in terms of resolution equivalence

[1:41:08 PM] DR: if you do a $get on (=bill/+email), do you get the same result as when you do a $get on =bill+email?

 

Giovanni said no.

 

Drummond then phrased the question this way:

 

[1:41:44 PM] DR: If you had the two XDI subjects (=bill/+email) and =bill+email in the same XDI document, are they synonyms by definition?

 

Again, Giovanni thought they would not be synonyms by definition.

 

Drummond then pointed out:

 

[1:44:35 PM] DR: #2 and #3 are XDI subjects - single XRI segments

[1:44:55 PM] DR: #1 is NOT an XDI subject - it is a full XDI statement consisting of subject, predicate, and object

 

Giovanni agreed. Giovanni then said:

 

[1:46:43 PM] GB: (=bill/+email) is our reification

[1:47:05 PM] GB: but a different subject than =bill+email

 

Drummond then asked:

 

[1:47:22 PM] DR: How is a reification of (=bill/+email) different =bill+email?

[1:53:00 PM] DR: My understanding has been that our definition of what =bill+email meant is that it was a synonym for the reification of =bill/+email, i.e., (=bill/+email), so that by definition they represent the same concept, whatever that concept is.

 

Giovanni then asked why XDI needed a synonym for (=bill/+email)?

 

[1:55:04 PM] DR: The motivation I believe has always been syntactical simplicity - in other words it is much easier to read and write =bill+email than (=bill/+email).

 

At this point they began talking about the English phrase “Bill’s email”. Drummond said that “Bill’s email” is different than “Bill email”. The latter is more like “wood desk” – it qualifies a kind of desk. “Bill email” qualifies a kind of email, but is not an instance of email. “Bill’s email” is an instance of email.

 

Giovanni said that he believes we need a semantic identifier for "Bill's email" that is different than "Bill email".

 

 [2:00:38 PM] GB: I'd like to have =bill+email as the concept identifying Bill's email, whereas I'd like to have (=bill/+email) as the subject reifying the statement =bill/$has/+email

 

Drummond felt they were getting to the heart of the issue, because the XDI dictionary and data sharing logic that he's been working is based on the definition that =bill+email is a synonym for (=bill/+email). He felt that the English language concept of “Bill's email” would be expressed by the XDI statement =bill/+email (note that there is no reification).

 

Giovanni then asked how you would say:

 

[2:02:37 PM] GB: Bill's email has a capability of 200Kb

 

Giovanni felt it would be:

 

[2:02:56 PM] GB: =bill+email/+size/"200Kb"

 

Drummond felt it would be:

 

[2:03:57 PM] DR: (=bill/+email/$$)/+size/"200Kb"

 

(Note: later, after the telecon, Drummond realized an easier way to say this – he will post a new message to the list about this.)

 

Drummond said that to talk about a particular instance of email, you have to have an actual predicate, not the reification of a subject/predicate relationship.

 

At that point Giovanni felt this might be a semantics problem. He started to explain it…

 

[2:05:44 PM] GB: =bill+email

[2:05:55 PM] GB: =bill+email/$is$a/+email

[2:06:06 PM] GB: but it is a particular email, the one Bill has

 

…but at that point they ran out of time on the call (Drummond had the XRI TC telecon). So they agreed to hold another 1-on-1 telecon tomorrow to try to finish the conversation and “get to the bottom” of $has semantics.