Re: rel:foo for those who can't do without 'relative' URIs

-----Original Message-----
From: Jonathan Borden <jborden@mediaone.net>
To: keshlam@us.ibm.com <keshlam@us.ibm.com>; xml-uri@w3.org <xml-uri@w3.org>
Date: Wednesday, June 07, 2000 10:07 AM
Subject: RE: rel:foo for those who can't do without 'relative' URIs


>keshlam@us.ibm.com wrote:
>
>>
>> >with this in place, is there *any* reason (besides the 3 legacy
documents
>> >:-) not to ban relative URI references as namespace names?
>>
>> I think that's always been the only serious objection to banning them.
>>
>> If the folks who felt they couldn't live without relative syntax (for
fear
>> of breaking legacy documents) are willing to take this solution,
>> GO FOR IT.
>> I've lost track of whether that's still a concern or not.
>>


Thatis my sense of the consesnus here.


>Assuming that we accept the replacement of "foo" with "rel:foo", would it
be
>acceptable to define a 'migration' path for XML namespace in the
deprecation
>of relative URI references to define the behavior of a parser when a
>relative URI reference is encountered to prepend "rel:" and proceed?


No, I think you just have to accept that documents which use relative
URI-references
may under rather obscure circumstances give unexpected results in a
well-fromedness test.

The whole point of the problem is that you can't tell by looking at a
document whether
a relative URI-reference was written as a relative-URI reference (reading
the first
part of the NS spec) or as a string (reading the second part).

>Thus a legacy XML document would be automatically transformed into a 'well
>structured'/namespace conforming XML document.
>
>Jonathan Borden
>
>

Received on Wednesday, 7 June 2000 22:55:46 UTC