bryan rasmussen wrote:

> On 10/17/06, Florent Georges wrote:

> > I don't say it is not correct to use matches() to check
> > the lexical representation of a value, in order to try
> > to deduce some information about its type.

> I was pointing out that while Roger may have phrased it
> that way the reference to schematron seems to indicate
> that what he actually means is that with matches he can
> constrain the value.

  Maybe.

>                      It is perhaps an arrogant assumption
> on my point but I think he means that he thought that
> matches, using Unicode Regex would more closely and easily
> match the requirements expressed in the schema because of
> basically using the same constraint 0-9.

  I can't see how using matches() can "more closely and
easily match the requirements expressed in the schema" than
a schema validation itself.  And as Mike pointed out, what
about " 1 "?

> >  But then you have to be aware you'll not get the type,
> > you'll get guesses.

> As stated earlier I don't believe he wants to get the
> type, he wants to constrain the value, and as for me I am
> not at all interested in the type.

  Maybe.

  Regards,

--drkm



















	

	
		
___________________________________________________________________________ 
Dicouvrez une nouvelle fagon d'obtenir des riponses ` toutes vos questions ! 
Demandez ` ceux qui savent sur Yahoo! Questions/Riponses
http://fr.answers.yahoo.com