Using "Id" attribute names and possible confusion with type ID attributes

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

Here is my promised investigation into the use of "id" and its variants as an attribute name and the potential confusion with attributes that are assumed to contained type ID values without checking a schema.

My conclusion is that there is no difficulty using "id" as the name for identification of a test assertion, especially if it is *not* an attribute whose value data type is ID.  There is therefore no interference with xml:id were it also used.  (xml:id is only XML document unique, however, and is not unique in any more-global sense).

 - Dennis

BACKGROUND

Tim Bray posted about the general problem of "id" attributes across various formats,
<http://www.textuality.com/xml/naked.html> of April, 1999 but this does not hinge on how attributes are named in XML.

The xml:id Version 1.0 W3C Recommendation 9 September 2005
<http://www.w3.org/TR/2005/REC-xml-id-20050909/>
provides xml:id as a solution to the problem of otherwise not knowing what are the attributes of type ID.
(It is interesting that all values of type ID in an XML document instance must be unique within the document, but the xml:id specification does not restrict the number of attributes of type ID that an element can have.  The XML specifications do restrict each element to at most one attribute of type ID.)

I thought there was a problem related to DTD compatibility in Relax NG but I can find no relevance to how ID appears in attribute names in Relax NG material.


-----Original Message-----
From: jdurand@us.fujitsu.com [mailto:jdurand@us.fujitsu.com] 
Sent: Tuesday, May 10, 2011 17:13
To: tag@lists.oasis-open.org
Subject: [tag] Groups - TestAssertionMarkup-CSD4-May10-diffs (testassertionmarkuplanguage-1.0-csd-04-May10-diffs.odt) uploaded

New candidate CSD04 TA Markup (rev May/10) , with diffs from previous:
[ ... ]
- pending: renaming of testAssertion/@id as @taid or @tid.

 -- Mr Jacques Durand