a proposition about ID/IDREF and element optionality

A friend of mine was recently confronted with a curious claim
about XSD schemas; let us call it proposition (1):

     ... a schema that supports id/idref use must make all
     attributes and elements optional...

This seemed implausible to them, so they consulted me, just to
make sure there was not some sub-clause of a sub-clause they had
overlooked.

I have assured them that unless there is a sub-clause of a
sub-clause which I have also overlooked, proposition (1) is
false.  Now, it's not unheard of for people to make mistakes
about XSD and for false statements to circulate as a result.
But what puzzles me is that I can't think of any true statement
about XSD that might have been misunderstood in such a way
as to give rise to proposition (1).

Have any readers of this list encountered this misconception
before?  Do you know how it arose?

Have any of the various "best-practices" documents which convey
advice for users of XSD suggested such a thing?  Is there
some design discipline for XSD which prescribes optionality
for all things as the concomitant of ID/IDREF use (and if so,
what sects practice that discipline)?

And, of course, if you can point me to the sub-clause of a
sub-clause which makes proposition (1) true, I am eager to
be enlightened, even at the cost of public humiliation.

Thanks.

-- 
****************************************************************
* C. M. Sperberg-McQueen, Black Mesa Technologies LLC
* http://www.blackmesatech.com
* http://cmsmcq.com/mib
* http://balisage.net
****************************************************************

Received on Saturday, 6 February 2010 01:04:29 UTC