RE: 4. ID assignment and the empty string

> From: public-xml-id-request@w3.org  On Behalf Of Chris Lilley
> Sent: Wednesday, 02 February, 2005 22:38
> To: Ian Hickson
> Cc: Norman Walsh; public-xml-id@w3.org
> Subject: Re: 4. ID assignment and the empty string
> 
> 
> On Wednesday, February 2, 2005, 3:45:44 PM, Ian wrote:
> 
> 
> IH> On Wed, 2 Feb 2005, Norman Walsh wrote:
> >> 
> >> 1. Rather than speaking of "ID assignment", the specification now
> >> speaks of "ID type assignment": [...]
> >> 
> >> 2. We added a note to make it clear that application 
> behavior (e.g.,
> >> whether or not the getElementById() function actually accepts the
> >> empty string as a legitimate value) is beyond the scope of this
> >> specification. [...]
> >> 
> >> Please let us know if this change satisfies your comment. (Our CR 
> >> decision call is tomorrow morning, so a prompt reply would be most 
> >> appreciated.)
> 
> IH> This change does satisfy my concern, thanks!
> 
> Leaving something deliberately unspecified is one way to proceed, but
> not a way that I like.

While I appreciate that, the xml:id spec is not trying
to solve all the interoperability problems with all 
XML related tools.  No one has given the XML Core WG
charter to decide how, say, browsers should work.

Regardless, we tried hard to have what xml:id says
match what's true about XML 1.0 attributes declared
as ID.  While we tried in some cases to add some
clarity to the issue in xml:id where possible, we
are basically trying to say that an attribute named
xml:id can be used to accomplish many of the good
things an ID attribute can but without the need
for any declaration.  

As it stands, in XML 1.0, an attribute whose declared
type is ID can be assigned the null value in 
well-formed XML, yet I understand that browsers will
not assign such an id value to an element.  So the
same situation already exists for IDs as we have now
allowed for xml:id.

paul

Received on Thursday, 3 February 2005 15:09:06 UTC