> <rant> > The cause of this problem is the short-sighted and idiotic W3C > decision that > allowing a cardinality of other than one in an xs:all construct would be > "too difficult for processors to implement." This was such a stupid move > that it had to be political in nature and makes me wonder which member's > processor was so poorly written that they couldn't handle implementing > something like: > > <xs:all> > <xs:element name="n" minOccurs="3" maxOccurs="5"/> > ... > </xs:all> > > That is not a difficult case to handle at all. A mere few moments of > thought reveals at least one reasonably efficient and recursable > algorithm, > there are probably many others. > </rant>
Couldn't be further from the truth. Certain classes of validators based on finite state machines increase in complexity exponentially when cardinalities are not restricted. The only "political" thing is that this was a compromise position by thoughs who wanted to exclude the feature completely and those that wanted no restrictions. If you want to debate technical merits, go to xmlschema-dev where there are plenty of people knowledgable with this issue. David Cleary --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
