On Wed, 18 Aug 2004 12:16:07 -0400
Norman Walsh <> wrote:
> What should we do about XML 1.1?
>
> I propose an erratum that says processors may support XML 1.1
> characters in names.
Let us consider XML 1.1 RNG schemas and XML 1.1 instances separately.
1. XML 1.1 RNG schemas
As for XML 1.1 RNG schemas, I have a mixed feeling. XML 1.1 allows more name
characters, some of which are strongly needed by somebody (including
myself). However, I have some concerns.
First, to support XML 1.1, RNG implementations have to
(1) allow more name characters to occur in schemas,
(2) trim leading or trailing #x85 and #x2028 in tokens occurring in schemas, and
(3) allow C0 control functions to occur in schemas
I guess that (3) does not require any changes to implementations.
(1) and (2) do require changes.
Second, even if we do the above changes, the datatype library for W3C
XML Schema supports XML 1.0 name characters only. It is not clear when
W3C XML Schema 1.1 appears.
Third, it is not clear whether other parts of DSDL support XML 1.1 name
characters. I plan to raise this issue in the SC34 WG1 meeting in
September.
2. XML 1.1 instances
We can count on XML 1.1 processors for handling XML 1.1 instances.
I guess that Jing and MSV already support XML 1.1 instances, since
they use Xerces-J. Thus, I do not see any technical problems.
My only concern is alignment with the other parts of DSDL. Please
wait for the resolutions at the WG1 meeting in September.
Cheers,
--
MURATA Makoto (FAMILY Given) <>
