[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: XDI metagraph model, natural language examples, and a stunning result
I had the action item from the telecon last week to provide some clear examples of the natural language semantics that I believe the XDI 1.0 metagraph model needs to be able to support.
I will provide a few examples, and then show a stunning conclusion about one that we have been discussing for several months now.
EXAMPLE #1: NOUNS SPECIALING NOUNS
This example is when one noun (RDF subject) is used to specialize another noun to form a new composite noun (i.e., new RDF subject).
car engine
boat engine
airplane engine cowling hinge screw
home phone
work phone
emergency day phone
emergency night phone
In essence the concept is expressing hierarchy by expressing the relationship between a series of nouns (RDF subjects).
Note that this same principle extends to adjectives describing nouns.
fast engine
slow engine
old greasy leaking engine
EXAMPLE #2: POSSESSIVE NOUNS
In English, this is the use of a possessive (http://en.wikipedia.org/wiki/Possession_%28linguistics%29).
Bill’s car
Bob’s car
Jack’s friend’s car
EXAMPLE #3: PLURAL NOUNS
This is the form of a noun that refers to more than one referent (http://en.wikipedia.org/wiki/Plural).
cars
engines
car engines
fast car engines
Bill’s cars
PROPOSAL FOR EXPRESSING EACH EXAMPLE WITH THE METAGRAPH MODEL
In the current XDI 1.0 metagraph model, example #1 above (specialization of a subject with another subject) is defined as a $has relation between the two subjects.
+a/$has/+b <==> (+a/+b) <==> +a+b
+car/$has/+engine <==> (+car/+engine) <==> +car+engine
However to date we have not identified a satisfactory expression for example #2 (possession) or example #3 (plurality). In examining the metagraph model over the past week, based on observations from Giovanni on last week’s telecon, I discovered a pattern that appears to have the correct semantics. What will stun you is what the final expression looks like in XDI.
The pattern is one of expressing the relation of an XDI RDF subject to its own context, which in the metagraph model is expressed with a single dollar sign, $. In the case of the subject =bill, this means:
=bill/$has/$ <==> (=bill/$) <==> =bill$
By the semantics of the metagraph model, this is expressing the relationship that “Bill has (possesses) a context”.
To then express that another noun (RDF subject) is contained in (addressable in) this context, we use another $has statement:
=bill$/$has/+car <==> (=bill$/+car) <==> =bill$+car
From a metagraph perspective, this literally means “the RDF subject +car appears in the XDI context possessed by =bill”.
As if that is not fascinating enough, here’s how the same approach applies to expressing example #3, plurality. If the XDI RDF subject is not an instance (like =bill), but a class (like +car), then the same pattern can be used to express that all the subjects in a particular XDI context are members of that class.
+car/$has/$ <==> (+car/$) <==> +car$
From a metagraph perspective, this literally identifies “the XDI context possessed by +car”, all of whose XDI RDF subjects must be members of the class +car. In English, this is the plural form of the word, i.e., “cars”.
Put this all together, and you can express both possession by an instance (“Bill’s”) and the plural form of a class (“cars”) using a composite $has statement:
=bill$/$has/+car$ <==> (=bill$/+car$) <==> =bill$+car$
I look forward to discussing on today's call.
=Drummond
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]