[Top] [All Lists]

Re: [ontolog-forum] Semantic Systems

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sun, 28 Jun 2009 08:33:32 -0400
Message-id: <4A47631C.3080306@xxxxxxxxxxx>
Duane,    (01)

DN> JSON is also specific to a certain runtime.
 > Is KIF not the most prominent standard in this area?    (02)

Those two points require a considerable amount of
clarification and qualification.    (03)

JSON is a notation that is used in Javascript.  But it is
a notation for data, which could be processed by any kind of
programming language for any purpose.    (04)

Check the opening page of http://json.org which describes the
complete grammar in 5 rules (object, array, value, string, number).
It also has mappings to 39 different languages, ranging from ASP
to Visual FoxPro and almost anything you can think of in between.    (05)

There is also another page that describes JSON as "the fat free
alternative to XML":  http://www.json.org/fatfree.html    (06)

However, you are correct in saying that the semantics of JSON
depends on the runtime of the language into which it is imported.    (07)

JSON could be used as an alternative to XML for the syntax, but
the lack of a standard semantics for JSON means that it could
*not* be used as a replacement for RDF *unless* an official
standard were adopted for mapping RDF to and from a particular
subset of JSON whose semantics was defined in Common Logic.    (08)

However, that subset of JSON could also be generalized to include
an official standard mapping of any table of a relational database
(or in fact *all* the tables of a DB) to and from JSON.    (09)

Re KIF:  The Knowledge Interchange Format (KIF) is the predecessor
of CLIF.  Back in the 1990s, I started working with standards
committees (originally X3 H4, then NCITS T2, then INCITS L8),
which were the US TAGs to ISO/IEC JTC1 SC32 WG2.    (010)

I spoke with Mike Genesereth and Richard Fikes (the coauthors
of the KIF report) about developing parallel ANSI standards for
KIF and Conceptual Graphs -- i.e., define a common semantics for
both with translations between the two.  And they agreed.    (011)

After a lot of work that took a long time (primarily because Mike
and I kept getting distracted with other projects related to our
day jobs), the goal of parallel ANSI standards became parallel
ISO standards, and that goal was replaced with a more streamlined
goal of a project called Common Logic with one ISO standard for
a common semantics and two alternate syntaxes (KIF and CGIF).    (012)

But then the Semantic Web guys developed two versions of logic,
which they called RDF and OWL, and the common semantics for KIF
and CGIF was not sufficiently general to support it.  So a couple
of other guys, Pat Hayes and Chris Menzel, suggested a brand new,
more general semantics for Common Logic.  The advantage of the
new semantics is that it could also support RDF and OWL.    (013)

The new semantics was almost upward compatible with KIF, but not
quite.  There are some KIF specifications that require a bit of
tweaking to map to the new notation.  Because of that tweaking,
the ISO WG chose the name CLIF rather than KIF2.    (014)

Getting back to JSON:  Since JSON has no semantics, it would be
better to map RDF to Common Logic (which has a semantics that
can support RDF).  Then that mapping could be represented in
both CLIF and CGIF.  That subset of Common Logic could be
called RDF-CL.  But RDF-CL could also be represented in a
new dialect for a subset of Common Logic that just happened
to look like JSON.  That dialect could be called RDF-JSON.    (015)

And a further generalization of RDF-CL could also represent
any table in Common Logic.  That representation with the
semantics defined by CL could map any DB table to CLIF or
CGIF or a JSON array whose members happened to be JSON objects.    (016)

Some qualifications must be added to accommodate the open or
closed world assumptions, but those assumptions concern the
way those tables are *used* in inferences.  The raw data
in the table would be represented in as a JSON array of
objects -- but with the assumption that certain CL axioms
determine the meaning of that array of objects.    (017)

The open and closed world assumptions would determine
different CL axioms about that data.    (018)

This is a rather long answer to two very short questions,
but that is life.    (019)

John    (020)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (021)

<Prev in Thread] Current Thread [Next in Thread>