Proposed schema issue: clarify definition of "restriction"

Our recommendation in chapter 2 introduces the definition of restriction as
[1,2]

"[Definition:]  A type definition whose declarations or facets are in a
one-to-one relation with those of another specified type definition, with
each in turn restricting the possibilities of the one it corresponds to, is
said to be a restriction. The specific restrictions might include narrowed
ranges or reduced alternatives. Members of a type, A, whose definition is a
·restriction· of the definition of another type, B, are always members of
type B as well."

First of all, I note that the first two sentences of this definition well
likely emerge as too narrow if we adopt a more generalized "restriction is
subsumption" rule for Schema 1.1.  That, however, is not the issue raised
by this note.

It is my belief that the last sentence is crucially important to the
recommendation, and is the place where we (attempt to) say normatively:
"restriction is at least subsumption".  In other words, we may at times
rule out even certain possible that are in fact subsumptions, but by
definition we can never have a restriction that violates subsumption.  More
informally, if any other part of the rec appears to provide for
construction of a "restriction" that allows content not allowed by the
base, that is ipso facto a contradiction in the recommendation and should
result in an erratum.

In today's discussion at the Cambridge F2F I was told that the word
"member" in the sentence in question is deemed by some to be ambiguous or
unclear, and the sentence is therefore not taken to have its intended force
of ensuring that "restriction is at least subsumption".  Furthermore, I was
informed that earlier discussions (which I do not specifically remember) of
this terminology have proven diffficult.   Some of the difficulties may
relate to the distinction between lexical and value spaces.   Nonetheless,
I hereby request that we open an issue to clarify this sentence, and
strongly urge that it be written to clearly state that restriction is
indeed at least subsumption.

Furthermore, I would suggest that this is an urgent problem, and that we
should attempt to resolve it in our next recommendation-level release,
which might be an erratum to Schema 1.0 or might be Schema 1.1.

Noah

[1] http://www.w3.org/TR/xmlschema-1/#key-typeRestriction
[2] http://www.w3.org/TR/2004/PER-xmlschema-1-20040318/#key-typeRestriction

--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Thursday, 13 May 2004 15:14:56 UTC