[Om3] Pragmatics, timing, summary, and proposed resolutions

[David Carlisle]

> Is it not possible to make SCM's alignment with both OM-n and maybe other
> non-OM formalisations 'dynamic'?

yes of course but at a cost.
We could (taking it to extremes) declare the content type of every
element to be open and any attributes to be allowed, and just state
that MathML defines the semantics for a range of expressions and that
any other expressions are nt invalid but are curently undefined in mathml3.

This allows arbitrary future extension but gives up on any error
checking or schema context driven editing help on current documents.

In practice of course you end up doing someting in between and making
each construct open or closed depending on its merits as you see it, but
the tradeoffs are not really so different in each case.
But the OM design really is for a fixed stable set of primitives
with exteibility being by the CD mechanism specifically so it doesn't
require XML level schema changes to extend. That's not to say that OM
shouldn't be extended occasionally but that should be rather more rare
than most XML vocabularies.

In particular every MathML exression is supposed to have a default
rendering so the more you open up the schema the more you have to
specify (or wave hands to avoid specifying) what some arbitrary
construct is supposed to look like.

David

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 
________________________________________________________________________