Re: 3. Frequency and timing of xml:id processing

/ Ian Hickson <ian@hixie.ch> was heard to say:
| On Wed, 5 Jan 2005, Norman Walsh wrote:
[...]
|> I'm inclined to say that this processing occurs whenever the xml:id
|> Processor is invoked and that it is up to the application to decide
|> when this occurs.
|
| Could some requirements be placed on certain classes of applications?

I don't know of any broadly accepted classification scheme for all
existing and yet-to-be-invented applications. Without such a scheme, I
don't know how the xml:id specification could identify classes of
applications in any practical way. I fear that attempting to classify
existing applications without some way to classify the ones that
haven't been invented yet will just make xml:id become more confusing
as time passes.

| It 
| would be unfortunate if one Web browser never updated its concept of 
| IDness where another maintained its IDness live with document changes, as 
| the results for scripts, stylesheets, XForms, et al, would be radically 
| different. If both behaviours are conformant, then interoperability will 
| be hard to obtain.

This problem already exists. At the moment ID assignment is only
normatively performed by the parser. I believe that our statement,
"Users of applications that provide facilities for modifying XML
documents may reasonably expect xml:id processing to occur whenever a
change is made to an ID value", will encourage interoperability.

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.