Re: Test case regarding XML Literals and octets

>Hi Pat,
>
>ok, I hadn't understood that you had chosen xml-exc-c14n in order to 
>get an XML representation minimally affected by the omitted XML 
>context. Sorry about this nonunderstanding.
>
>One more question (sorry if it is another uninformed one...): You 
>already specify (in the editors' draft) that the lexical form of XML 
>literals must conform to xml-exc-c14n (which I read to say, it must 
>be exclusively canonicalized XML).
>
>Since Exc Canonical XML is the serialization of a node-set, if you 
>just parsed that canonical XML you would get a node-set which is 
>exactly in the form you want: Minimally affected by the omitted XML 
>context. Wouldn't it make sense to specify that the denotation of an 
>XML literal is that node-set? Wouldn't that meet your goals 
>(minimally affected by context) and still match the "intuitive" 
>expectation that XML "denotes" an abstract tree?

As  far as I understand it, yes.

>I.e. I think your concerns are taken care of by picking xml-exc-c14n 
>on the lexical space level, the decision on the value space level 
>can in principle be made independently of that? Or am I missing 
>something simple again?

No. However at this point I am speaking slightly out of turn, as the 
details of the XML literal definition were put together by Jeremy 
Carroll and he may have his own reasons for preferring to be explicit 
about octets (as indeed may other members of the WG). I should 
perhaps have made it clear that I was speaking as the semantics 
editor, and as far as that goes the two key points are that XML 
literals should be exactly 1:1 with the literal character strings and 
also if possible be entities which are distinct from character 
strings (or else, possibly, identical to character strings; but in 
either case, clearly all one way or the other, i.e. it would not be 
workable to have, say, XML literals without markup denote character 
strings but XML literals with markup denote something else.)

>pat hayes wrote:
>[...]
>>The XML exclusive canonicalization document itself describes the 
>>purpose of exclusive canonicalization in these terms, as being the 
>>reason for introducing it in the first place.
>>(section 1.2 Applications:
>>"The applications of Exclusive XML Canonicalization are very 
>>similar to those for Canonical XML [ XML-C14N ]. However, exclusive 
>>canonicalization, or equivalent means of excluding most XML 
>>context, is necessary for signature applications where the XML 
>>context of signed XML will change ")
>[...]
>>  The fact that
>>the document refers the definition of this form of canonicalization 
>>to octets rather than nodesets (which does seem odd, now you point 
>>it out)
>
>Well, it says that the applications for xml-exc-c14n are similar to 
>those for xml-c14n-- i.e., getting a canonical *representation* of 
>an XML document. That xml-exc-c14n isn't defined in terms of a 
>nodeset can probably simply explained by the authors' being the XML 
>Signature Group and mostly concerned with signatures.

Ah, I hadn't thought that aspect through, I have to admit.

Pat


-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Friday, 1 August 2003 15:24:30 UTC