Defining the MCE smentics

[MURATA Makoto]

Dear colleagues,

I propose to use both a declarative approach and a procedural approach
for defining the MCE semantics.

1) Declarative approach

In this approach, we define which element or attribute in a given
document is retained and which is discarded by providing a set of
declarative positions.  These propositions are collected in one place
in the MCE specification rather than being scattered in many clauses.

Every thing about the MCE semantics should be clarified by
these propositions.  The only possible exception is timing
of error handling such as non-understood @MustUnderstand.

Examples of propositions are shown below:

       An application-defined extension element as well as its
       attributes and contents is retained unless some ancestor
       element is ignored or unselected,

       A Choice element is selected if and only if the required
       namespaces are understood and no
       elder sibling elements are selected.  It is unselected
       otherwise.

       A Fallback element is selected if no
       elder sibling elements are selected.  It is unselected
       otherwise.

       Unselected Choice or Fallback elements as well as
       their attributes and contents are discarded.
'
Examples should clarify tricky cases (such as ignorable
application-defined elements).  I expect about ten examples.

2) Procedural approach

In this approach, we provide a non-normative procodure for illustrating
the behaviour of the MCE eprocessor.

We might want to use an XSLT stylesheet as shown in my previous mail.
Another possibility is to provide a procedural algorithm in some
programmnig language or prose, but I think that XSLT is compact
and comprehensive for MCE semantics.

This procedure has nothing more than the normative propositions
with the possible exception of timing of error handling.

3)  Reference model

I now feel that the declarative propositions and procedure (as well as
examples) sketched above are good enough for clarifying the
semantics of MCE.  I do not see strong reasons for a clause
dedicated to the processing model.


--

Praying for the victims of the Japan Tohoku earthquake

Makoto


-- 

Praying for the victims of the Japan Tohoku earthquake

Makoto
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.vse.cz/pipermail/sc34wg4/attachments/20120425/66377433/attachment.htm>