RE: multiple inheritance
Christopher,
XML is an interesting case since it is being employed heavily to support
procedural programming languages. The Java-XML connection is so strong, it
is reasonable to assert that it's advantageous for their data models to be
similar. At the same time, those in the knowledge representation community
are likely to look at XML and wish for many of the features that are common
in KR languages.
I wonder though about Henry's incremental approach. It's fine to say
that XML itself will evolve in steps from single to multiple inheritance
(and we might wonder if it will continue to accept new features, expanding
its expressiveness), however, that may have a catastrophic effect on
*content* standards that use XML as their syntax.
For any language, I believe that there are expressions within each
language, apart from correctness, which are "idiomatic" or natural, and
expressions which are not. This is similar to a non-native speaker of a
human language who may utter sentences which are grammatical yet sound
unnatural to a native. Any mature C++ programmer can see when a C
programmer is just using C++ as a C with more features who has not grasped
how to write idiomatic C++ code for example.
I would suggest that if we were to create an ontology assuming single
inheritance that it would be prohibitive to re-engineer the standard later
on to support multiple inheritance. It would be either a completely
different ontology or one which is not idiomatic for the new language.
Adam
ref: http://ltsc.ieee.org/logs/suo/msg00025.html
At 04:31 PM 6/16/2000 +0200, Christopher Spottiswoode wrote:
>Members of this list who are not on the XML.com list may appreciate being
>pointed to this paragraph from an article pointed to by the latest number
>from
>XML.com. (The article is at
>http://xml.com/pub/2000/06/xmleurope/schemas.html )
>
>The article reports on Henry Thompson, introduced as "an editor of the W3C
>XML Schemas specification", and is re his address, entitled "Reconstructing
>DTD Best Practice", at this week's XML Europe 2000 conference in Paris.
>
>"Thompson also explained that XML Schemas 1.0 would not include multiple
>inheritance, as the Working Group is keen to produce a strong design for a
>single inheritance model first. However, Thompson did not rule out that
>multiple inheritance could feature in a later version, and that a model
>similar
>to Java (single inheritance supplemented by interfaces) could still be
>considered at that time. The emphasis is clearly on getting the first
>version of
>XML Schemas complete."
>
>While I myself find it difficult to imagine a congenial conceptual world
>without
>multiple inheritance, I do find it interesting to ponder the W3C WG's
>strategies, where they come from, and what implications such work from
>such a massively influential stable will have on the practical future of
>ontologies.
>
>A next question might be what features a first-pass single-inheritance
>ontology should have if it is to grow, in due course, into multiple
>inheritance,
>and without enormous inconvenience. (It is reportedly aiming to be
>"complete", and presumably it aims to be open too.)
>
>Christopher.
>
-----------------
Adam Pease
Teknowledge
(650) 424-0500 x571