W3C home > Mailing lists > Public > www-tag@w3.org > June 2003

xmlIDSemantics-32

From: james anderson <james.anderson@setf.de>
Date: Tue, 17 Jun 2003 15:45:52 +0200
Message-ID: <3EEF1AEA.69759259@setf.de>
To: www-tag@w3.org


in the hope that, should xmlIDSemantics-32-20030512 become the basis for
future guidelines for processing xml, it might do so in a more carefully
reasoned form, i suggest that one consider:



1. the passages

   "IDness is a consequence of parsing a DTD, not of validation."
and
   "The concept of IDness, which exists in well formed documents,
    should be distinguished ..."

would benefit from a proximate definition of "IDness". especially if the
reader is immediately called upon to distinguish "IDness" from other concepts
and processes in terms of which it can be defined.



2. the claim

  "... these [ID validity] constraints only apply in the case of
   a validatable and validated document." 

is neither clear nor self-evident. [as an aside, i read the claim to be that
the "constraints apply in the case ... only."]

what does "validatable" mean? if a document is "not validatable" does that
mean that it can be inferred that "the document could not possibly be valid",
or does it mean that "it cannot be determined whether the document is valid or
invalid?" that it is possible to determine uneqivocally, that the example
document is invalid indicates that there are nuances to "not validatable"
which would bear explication.

that a document is valid does mean that the constraints apply to that
document. that does not mean that other documents may not satisfy the constraints.

in addition, even non-validating xml processors are required to process
aspects of an internal dtd subset. one of those aspects is attribute
declarations - at least to the extent that they establish constraints for
attribute value normalization. which processing establishes the logical
possibility of applying constraints of the same order as the cited validity
constraints - even if they are termed "IDness constraints" instead.



3. the meaning of the passage

   "Some specs assume that all documents with DTDs have IDs (untrue),
    or that parsers only fetch the DTD for HTML but not XML, or that
    the implementation has some 'built in' knowledge of particular
    namespaces and the attributes associated with them that are of
    type ID."

is not clear. i surmise that it was intended to mean something close to

"Some specs assume that all documents with [IDs have DTDs (a)] (untrue), or
that parsers [interpret the DTD for HTML only, but do not do so for XML-based
document types, and (b)] that the implementation has some 'built in' knowledge
of particular namespaces and the attributes associated with them that are of
type ID."

a: it is not clear how the stated assumption applies to the topic. another
reading of the original phrase corresponds to the observations in the
paragraph which follows in the text, and could be expressed as "any attribute
named ID in a document with a DTD is interpreted as having IDness."
b: the limitation is on the document type not on the action.



4. most of the document discusses dtd parsing, dtd processing, and dtd-based
validation as if they were synonymous. if it were more careful in this regard,
it would be possible to discuss "solutions" in addition to those which depend
on validation. there is a passage towards the end of the discussion which
introduces such distinctions as a basis for document classifications
alternative to "well formed" and "valid". a more complete version of that
passage would be much more useful towards the beginning of the document than
the present passage is at the tail end of the discussion. 



5. the passage

    "Requiring DTD validation to get IDs is too big a retrogressive
     step; it essentially throws away well formedness as a concept
     and also XML namespaces, and needlessly conflates validation
     with decoration."

warrants a much more thorough foregoing foundation than that which this
document provides. particularly under a w3c masthead. in what sense is a
process worse off if it ascertains the "ID properties" of an XML-1.0 document
by examining a DTD. worse off than what? as the document is not clear to
distinguish dtd parsing/processing/validation, the intend of the judgement is
not clear, but of one reading of REC-xml-20001006 is, such processing is
exactly what a conforming processor should be able to do.  independent of
whether it determines that the document is valid. independent of whether it
even attempts to determine whether the document is valid.

....
Received on Tuesday, 17 June 2003 09:39:38 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:38 UTC