W3C home > Mailing lists > Public > xmlschema-dev@w3.org > April 2008

Re: UPA or not UPA violation?

From: C. M. Sperberg-McQueen <cmsmcq@acm.org>
Date: Wed, 16 Apr 2008 11:53:41 -0600
Message-Id: <FF4C9654-C0E7-42DF-8C39-B8172879C042@acm.org>
Cc: "C. M. Sperberg-McQueen" <cmsmcq@acm.org>, xmlschema-dev@w3.org
To: lists@fgeorges.org


On 16 Apr 2008, at 05:52 , Florent Georges wrote:
>
> Michael Kay wrote:
>
>> Remember what UPA stands for: unique particle
>> attribution. Every child element can be unambiguously
>> associated with exactly one particle in the content
>> model. If there are no child elements, this is trivially
>> true.
>
>   Oh yes, I see my fault: I thought the processor had to know
> wether group was used, but the UPA doesn't apply at this
> level.  Thank you for the explanation.

Thank you for raising the question.  If it's helpful, you might
like to know that the difference between the two conceptions
of the rule seems to be more or less the same as the difference
drawn by some computer scientists between "weak" determinism
(which corresponds to the UPA rule: you always know which primitive
token in the regular expression matched the input) and "strong"
determinism (you know not only which particle you are in, but
also which repetition operators fire when.

If the PSVI associated with an element the annotations on the
groups in its content model which are satisfied, then the
ambiguity (not just non-determinism) of your example content
model would be a practical problem. Fortunately, no such
association is provided.

Michael Sperberg-McQueen
Received on Wednesday, 16 April 2008 17:54:15 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:15:46 UTC