Re: correct behaviour for entities as system identifiers

[Warren Hedley <w.hedley@auckland.ac.nz>]

Hi again folks,

Thanks for everyone's comments.

I should have made it clear that when I intend to validate a CellML
document against the DTD, I set the "use_mathml_dtd" parameter to
INCLUDE. Xerces works fine with this set to IGNORE, but nsgmls which
I use for most validation gives me the usual "element <math> not
defined" style errors.

Thanks for the confirmation that the error is in Xerces and not in
the DTD.

Patrick, assuming that David C is correct, I don't see why the two
versions you suggest in your 3rd point would differ. I should probably
mention I'm both the editor *and* author of the CellML spec :-).

Cheers,
Warren Hedley

-- 
Warren Hedley