Re: Meaning of normative references [was: Update on namespaces]

[Taken as read: fulsome commendation to *both* Len and Dan for effective
use of rhetoric.]

> Instead, we should have a note that says that every valid XML document
> will ipso facto conform to ISO 8879, and possibly refer the interested
       ^^^^^^^^^^
> reader to that document.

Meaning: "By that very fact" (as opposed to ipso jure "by the operation of
the law itself", says the OED.)

But how is this "fact" to be tested? (It's like saying "every bottle of
beer ipso facto contains 4% alchohol", when there's no way of actually
testing specific gravity for a reality check on any given bottle.) 

Ipso facto validity will come down, in practice, to validation by the XML
parser at hand. ("Feels like 4% to me!") 

> cannot be a normative requirement for XML conformance.  Rather, the
> XML spec must stand alone, but the resulting valid XML documents must
                                                                   ^^^^
> always be valid SGML documents.  So any reference to SGML in the XML
> spec must be for interest, not one that forces people to read ISO 8879:1986.

Again, how "must" (and why not "shall"?)

Would it be possible to move to "ipso jure" xml validity by introducing
the concept that SGML is "operationally normative" with respect to XML?

By "operationally normative" I mean: a free and technically excellent SGML
parser (say, nsgmls)  is set up at a sensible site (say ...). The location
of the parser and the command line to supply it are published in the XML
specification. Then, if someone (say, a contractor who is obligated to
produce SGML-valid XML, or the grad student at the end of his/her one-week
trek) can feed that parser the XML in question, and if there are no
errors, the XML is SGML.

Returning to the beer metaphor, under this proposal the person with the
bottle in hand doesn't have to read a physics textbook to test that the
beer is 4%. 

S.

P.S. Incidentally, Len is correct that if the XML specification does what
it says it does (that is, if the words "Valid XML documents are SGML
documents in the sense described in 8879" are true in reality as well as
intent), no one will ever have to read actually read the SGML standard,
because no need for clarification will ever arise, and the XML
specification will be able to stand alone. 

But that is asking humans (the ERB, us) to create a perfect specification
-- not possible in this world, no matter how well anyone "does their job." 
Therefore, some method of allowing SGML to be normative ipso facto and
ipso jure is necessary that does not require the standard to be read is
needed: why I propose the concept of "operationally normative."

Received on Tuesday, 24 June 1997 22:56:41 UTC