Re: [xml-dev] Caution using XML Schema backward- or forward-compatibilit

["Stephen Green" <stephengreenubl@gmail.com>]

Interesting that a semantic change can be made by designers/owners
of a schema/vocabulary but equally by implementers, but when making
a change, whether or not it falls within a particular concept of compatible,
the change maker may have no idea whether the change will be successful
and beneficial; yet one thing is likely - there is no guarantee of success
just because one uses a particular versioning strategy. The versioning
strategy is just one of two key aspects of a contract between designer
and implemeter (or implementer and implementer) - one is the schema
itself and the other is the versioning strategy or policy. So the policy and
schema are key constraints, perhaps along with rules about semantics.
However the other key factor is experimentation within those constraints
and trial and error regarding the effects of the changes made. This just
involves proper risk assessment and risk management and due process
regarding feedback, etc.

On another note, will the semantic web and 'Web 3.0' help with risk
management by firming up the meanings in the vocabularies? E.g.
would it be better to use RDF instead of XML Schema for a vocabulary
especially sensitive to semantic drift such as business documents?
Or if the schema is already set in stone, would it alleviate adverse
semantic drift to add an ontology, PSIs, hash purls, etc? I'm rather
hoping so. Then maybe a new schema could break structural compatibility
(perhaps by necessity and progress) but preserve semantic compatibility.

-- 
Stephen Green

Partner
SystML, http://www.systml.co.uk
Tel: +44 (0) 117 9541606

http://www.biblegateway.com/passage/?search=matthew+22:37 .. and voice