|
|
|
Donal Fellows: 05/02/2007 9:45 AM EDT
|
| |
Comment: |
Oh, and ID is supported directly by DOM implementations, whereas XPath/XPointer are not (usually).
|
| |
Action: |
Update
|
|
|
Donal Fellows: 05/02/2007 9:44 AM EDT
|
| |
Comment: |
I suppose we could always recommend that people use GUIDs for ID values, since they're unique in time and space. OTOH, they're also ridiculously long.
:-)
The other thing to note is that anything that can be done with IDs can be done with XPath/XPointer. It might not be at all elegant though.
|
| |
Action: |
Update
|
|
|
Michel Drescher: 05/02/2007 8:29 AM EDT
|
| |
Comment: |
I agree with Donal, that xsd:ID and xsd:IDREF are very useful data types.
However, it must be considered that xsd:IDs can lead to collisions when XML Documents containing xsd:IDs are composed with each other. For example,
consider a hypothetical workflow document structure that consists of several JSDL document templates. The probability of xsd:ID collisions is fairly
high, I presume.
We should carefully evaluate whether QNames as a naming and identification mechanism serves us much better.
|
| |
Action: |
Update
|
|
|
Donal Fellows: 05/02/2007 8:21 AM EDT
|
| |
Comment: |
Actual definition (without XSD baggage) is:
<xsd:attribute name="id" type="xsd:ID"/>
This would then be referred to from elsewhere via:
<xsd:attribute ref="id"/>
(Assuming I remember my XSD hacking right.)
|
| |
Action: |
Update
|
|
|
Donal Fellows: 05/02/2007 8:20 AM EDT
|
| |
Comment: |
Being able to support an id attribute on elements (which MUST be of type xsd:ID; name of attribute is less important) would be a strong bonus, and is
a feature that is relatively well-supported by document-oriented tooling (but not so well with object mapping, which can't be helped) as well as being
simple (but time-consuming; lots of identical changes) to do. I believe it should definitely be on our roadmap for the next version of JSDL that
permits a schema change.
I'm not certain that doing this via an extension is a good approach; maximum utility is only attained if it also constrains extensions (that define
new elements) to support the attribute.
|
| |
Action: |
Update
Priority changed from 0 to 2
|
|
|
|
|
Donal Fellows: 02/15/2006 10:49 AM EST
|
| |
Comment: |
Comment from Dave Snelling: naming elements is *very* useful for signing things with XML-Signature
|
| |
Action: |
Update
|
|
|
Ali Anjomshoaa: 02/15/2006 10:12 AM EST
|
| |
Action: |
Update
summary changed from Should we add Refs and names to the schema to Should we add referencing capabilities to the schema?
|
|
|
Ali Anjomshoaa: 02/15/2006 10:11 AM EST
|
| |
Comment: |
Require a normative element(s) for referencing outside dependencies such as WS-Agreement.
OR
JSDL template docs for multiple useage.
|
| |
Action: |
Update
|
|
|
Ali Anjomshoaa: 05/03/2005 11:39 AM EST
|
| |
Action: |
Update
assigned_to changed from 7964 to 100
group_artifact_id changed from JSDL version 1 to JSDL-post-v1
|
|
|
Stephen McGough: 03/15/2005 5:10 AM EST
|
| |
Comment: |
This needs to be resolved for JSDL 1.0 as it has impact on the structure of the schema if we want to prevent the user from having a ref and contents.
|
| |
Action: |
Update
|
|
|
|
