Re: Qnames in attrs and c14n

Hash: SHA1

/ Tim Berners-Lee <> was heard to say:
| Hmmm... either that screws up the use of qnames in attribute values
| even more, which makes any alternative xml-validatable RDF syntax look
| pretty longwinded with URIs,

I think long winded is ok. Especially if it works.

| Suppose we use single quotes around qnames and double quotes around
| everything else ;-)

I think you mean for attribute values that contain a single QName.
That's a small fraction of the problem space, though perhaps it's the
whole space for RDF serialization, I'm not sure.

A proposal, such as changing the quoting rules, that includes a
fundamental change to the XML parser (i.e., XML 2.0) might work,
technically. I doubt it works politically, but maybe it will someday.

In the short term, I think the right thing to do is discourage people
from using QNames where URIs would work.

An alternative, I suppose (thinking off the top of my head) would be
to allow authors to tell processors which prefixes have to be
immutable. If the author knows that the prefix "x" is used in a
context where the parser won't recognize it, she could use "XMLNS:"
instead of "xmlns:" to declare the prefix. But that degenerates to the
case where all prefixes are immutable...

                                        Be seeing you,

- -- 
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.
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <>


Received on Friday, 28 November 2003 14:41:21 UTC