Re: Contexts (was Classes vs. Instances)

["John F. Sowa" <sowa@BESTWEB.NET>]

Mick and Avril,

I'm combining my responses to your two notes because
they're closely related.  A unified architecture *must*
take into account all the technologies that are used
for software design and development.  Separating SQL from
the WWW is absolutely impossible and counterproductive.

MK> I totally agree that as an architectural picture you
 > would be nuts to use such a diagram as this. However i do
 > think that technology stacks have there place in giving
 > a "at a glance" idea of the basic technology dependencies.

I agree.  I have nothing against diagrams as a quick
overview of the big picture.  But the fact that the SemWeb
stack ignores SQL is a symptom that something is drastically
wrong with the foundation.

You can't have a unified architecture when the basic diagram
leaves out the database systems that drive the world's economy.

AS> Even though tens of thousands have lived very long with
 > SQL without types, moving into types would not be hard for them.
 > It is clear that a type hierarchy _editor_ for standard SQL
 > could be built yet not changing SQL in any way...

I agree that it should be done.

AS> Frame-based ontology editors can export many languages.
 > Do you have a clear idea of the syntax of some better language
 > than RDFS or OWL?

If you want a very readable graphic editor, I suggest that you
look at the Concept Maps from IHMC.  They have a version of
Cmaps that generates RDFS and OWL.

For a unified semantics that can support RDFS, OWL, SQL, and
many other things, I would suggest Common Logic as the
foundation, Cmaps or something similar as the graphic editor,
and import/export facilities to RDFS, OWL, SQL, and all
the Common Logic dialects, such as CLIF, CGIF, and XCL.

That would support a very wide range of declarative languages
with a common graphic front end and Common Logic back end.
There are a lot of other tools that should be added, but
such a framework could support them.

John