LANG: OWL Light and triples

My view of where we could go with the OWL Light abstract syntax is as
follows.

1. The heavy syntax is over triples.

2. We define an XML serialization for OWL Light that is as close as possible
to the abstract syntax. This is not an RDF/XML document.

3. We identify which heavy syntax corresponds to each construct in the light
syntax,

4. We identify the additional presentational information needed to round
trip the light syntax through the heavy syntax (e.g. that these particular
properties went together in this order in a frame on this class).

5. We develop an encoding of this additional presentational information in
triples.

6. We develop an XSLT transform from the light sytax into the full set of
heavy triples and additional presentational information triples.

With this we will have:

- a concrete light syntax suitable for hand-editing.
- the ability to turn that light syntax into a conformant RDF document,
which can be processed by RDF tools.
- the ability to mix and match light and heavy syntax within the RDF soup.
- the ability to not loose presentational information in the conversation to
triples.
- the ability to round-trip a light document through triples and back out
again to be the same light document.


(Note that the word "same" does not mean bitwise identical - there is a
value judgment involved).

Basically I am proposing that OWL Light is not an RDF/XML serialization but
is RDF, in the sense that it can be understood as an RDF graph.

Jeremy

Received on Friday, 22 March 2002 05:54:09 UTC