Re: Deterministic behavior in processors

Tim Bray (tbray@textuality.com)
Wed, 05 Aug 1998 07:02:47 -0700

At 12:09 AM 8/5/98 -0700, Joel Nava wrote:
>This idea of specifying the use of a validating processor over the
>wire got me to wonder what method(s) would be preferable for doing
>so. Any suggestions?

Now there's a good question. At the moment, we have no mechanism aside
from a prominent notice on the website/marketing-material/tech-doc saying

"output from the MegaSoft XML-frobnosticator v7.4 in general uses
external entities, and thus a validating processor is required to
process it"

or

"output from the MicroCorp XML-depullulizer v0.3 does not use
external entities and may be processed correctly by a non-validating
XML processor"

David Singer of IBM at one time suggested that we bless another
PI or something in the XML declaration, saying

<?xml version="1.0" extentities="no" ?>

which is an assertion that no external entities are used. -Tim