RE: Proposed restatement of syntax-based interoperability princip le ( was RE: Action item on syntax-based interoperability)

Dave, there are systems out there (witness X3D) that are interoperating 
nicely with multiple syntaxes because the standard on which they are 
based defines an abstract model first.  So one can say syntax is an easier 
foundation to use where one can get widespread agreement on one, 
but not that it is 'foundational' for the web. 
It's convenient.  Even that where syntax is convenient, 
it doesn't have to be one syntax as long as all syntaxes used 
within the model are publicly and clearly documented.

Security is an interesting corner case.

len

From: David Orchard [mailto:dorchard@bea.com]

+1.

While it is true that interoperability is only really achieved by the
fruitful union of syntax, APIs, data models, and semantics (heck add in test
suites...), syntax is the foundational piece.  Removing syntax leaves much
less chance of interop.  Granted, cases exist where interop has been
achieved without it, but there are far fewer of those.  In fact, a number of
security folks over wss land have been arguing that without syntax, there is
zero interop because security can only be applied on syntax.

Received on Saturday, 25 October 2003 15:30:12 UTC