[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Minutes: XDI TC Telecon Thursday 1-2PM PT 2010-04-15
Following are the minutes of the unofficial telecon of the XDI TC at:
Date: Thursday, 15 April 2010 USA
Time: 1:00PM - 2:00PM Pacific Time (21:00-22:00 UTC)
1) DISCUSSION/PROPOSAL FOR SOLVING "THE PROBLEM FORMERLY KNOWN AS $HAS SEMANTICS"
First, Giovanni went over the email he sent as an action item from two weeks ago:
He explained that of the two examples below, he preferred example #2 because using $1 as a predicate (e.g., as an index into an array) makes the most sense to him.
Giovanni didn’t think that the predicate +email$1 made sense in a global context because it was essentially proposing a instance of a class.
Drummond felt that +email$1 was the equivalent of saying “the first email”, i.e., it is a generic concept until it is applied to a specific subject, at which point it would mean, “=example’s first email”, and thus point to a specific instance.
Giovanni pointed out that as currently defined, +email$1 semantically means “+email/$has/$1”.
Drummond explained that this was what he now believed was the root of the “$has semantics problem”, which has been a persistent issue for the last year. In thinking about this problem over the last two weeks, he concluded the root of the issue is that for both practical graph inspection/resolution purposes, as well as semantic purposes, we need a clear, semantically unambiguous way to differentiate between:
1) Parent/child relationships between XDI subjects, e.g., +a+b
2) Subject/predicate relationships, e.g., +a/+b.
One way to do this is to dedicate $has semantics to expressing only one of the two relationships above, and to define a different mechanism for expressing the other one.
Since the core metagraph concept of $has is to describe an outgoing arc from a subject, Drummond proposed dedicating +a/$has/+b to expressing +a/+b. This means we need a different way to express parent-child relationships between XDI subjects. For that, Drummond proposed using XRI subsegment delimiters as predicates, with each delimiter being the predicate that expresses the relationship between the parent subject and all child subjects that use that delimiter type. For example:
+a/+/+b <==> +a+b
+a/$/$b <==> +a$b
+a/=/=b <==> +a=b
+a/@/@b <==> +a@b
+a/*/*b <==> +a*b
+a/!/!b <==> +a!b
Giovanni asked if +a/$has/+b <==> +a/+b still implied +a+b.
Drummond said he believed that +a/+b would still imply that +a+b is valid, and vice versa, however from a resolution standpoint, the presence of one in the graph would not imply the presence of the other. For example
does not imply that =example+email$1 exists in the same graph, and vice versa.
Drummond said that, while semantically linked, the two are separate statements. In English it is like the difference between a verb and the infinitive form of a very in English, e.g., "run" and "to run".
Joseph asked what would be the actual English translations of +a/+b and +a+b? Or to use more concrete examples, =drummond/+email and =drummond+email?
Drummond said =drummond/+email would be "Drummond has emails" and =drummond+email would be "Drummond's emails". In other words, the first expresses a subject-predicate relationship, while the second is simply a subject.
# DRUMMOND to send an email to the list with more details about this proposal.
2) X3J (X3 FOR JSON) SERIALIZATION FORMAT PROGRESS
We did not have time to go into this further but the priority still remains for this new JSON serialization format to be completely defined so that interoperable implementations can be built and tested.
3) NEXT CALL
The next call is next week at the regular time.