W3C home > Mailing lists > Public > www-forms@w3.org > August 2003

Clarification of id attribute

From: Daniel Fowler <daniel.fowler@focus-solutions.co.uk>
Date: Tue, 12 Aug 2003 11:10:29 +0100
Message-ID: <30A02A46CB77D511851900508BAEADBCD1F4FE@exchange.focus-internal.co.uk>
To: "'www-forms@w3.org'" <www-forms@w3.org>

Hi All,

I'd like some thoughts on the following issue, I sent a mail to the editors
on 27/6 so it was probably to late for the Proposed Recommendation.

The Proposed Recommendation refers to the use of "the terms may, must and
should in accord with [RFC 2119]". The last paragraph in section 3.2.1
states "A host language must include an attribute of type xsd:ID on each
XForms element."

However, the example "G.1 XForms in XHTML" does not include an attribute of
type xsd:ID on all the XForms elements.

Thus in the last paragraph in section 3.2.1 I believe that the word "must"
would be better replaced with the word "should". This is supported by the
wording in the section 2.4 which explains how 'scoped resolution' works for
when the binding element does not have a model attribute. And in such
circumstances the id attribute on a model is not neccessarily used ("the
first XForms Model in document order is used"). This and the fact that
XForms can be authored and work in current implementations without an id
attribute on XForms model elements suggests to me that under 3.2.1 it should
state "A host language should include an attribute of type xsd:ID on each
XForms element". (My understanding of the difference between "must" and
"should" comes from RFC 2119).

Or am I missing something, should EVERY XForms element have an attribute of
type xsd:ID?

Furthermore can someone please clarifying exactly how the attribute of type
xsd:ID should be declared. For example could a host language which defines
an attribute of type xsd:ID called "key" use this "key" attribute on the
XForms elements as suggested in the last paragraph in section 3.2.1.

There is also some inconsistency within the Proposed Recommendation with
regards to the id attribute so if anyone could also clarify it for me I
would be grateful. In several places the Proposed Recommendation uses ID, id
or xsd:ID. It is not always clear whether or not these occurrances are
refering to an attribute of type xsd:ID or an attribute called id of type

Here are the occurances of id, ID or xsd:ID that I have found that are not

a) Section 2.4 1st paragraph - id
b) Section 2.4 paragraph beginning "The main difference..." - ID
c) Section 3.2.1 last paragraph - xsd:ID
d) Section 3.2.3 in paragraph under "model" - ID
e) Section 3.2.3 in paragraph starting "It is an exception..." - ID
f) Section 3.2.4 in paragraph under "model" - ID
g) Section 3.2.4 in paragraph starting "It is an exception..." - id
h) Section 3.3.4 1st paragraph - xsd:ID
i) Section 4.5.1 1st paragraph - ID
j) Section 7.3.1 - ID
k) Section 7.5.1 paragraph starting "Another dynamic..." - xsd:ID
l) Section 9.3.2 last paragraph - id
m) Section 10.1.9 in the "setvalue with Expression" example - id

This should prove useful to those polishing up the Proposed Recommendation
before it reaches full Recommendation.


Daniel Fowler
Received on Tuesday, 12 August 2003 06:09:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:36:08 UTC