Re: Banning relative - No real damage?

At 11:45 AM 2000-06-07 -0400, Simon St.Laurent wrote:
>At 09:29 AM 6/7/00 -0400, keshlam@us.ibm.com wrote:
>>
>>> the "absolutize" folks are willing to put up with "forbid".
>>
>>If so, the challenge reduces to one of dealing with the legacy documents
>>already using relative syntax, which would become incorrect if forbid
>>isaccepted. We had a few folks strongly asserting that they absolutely
>>needed this, or at least needed some reasonable deprecation/migration path.
>>I've lost track of whether that's still the case.
>
>Because there is no mechanism at present for alerting processors to
>different versions of the namespaces spec, I'd suggest that the W3C start
>with deprecation, and only move on to full 'forbidden' status with another
>revision of XML.
>
>Basically, this will put out a warning (one I gave repeatedly at JavaOne
>yesterday) that relative URI references are dangerous practice and suggests
>that they will be excised from the spec, while letting current processors
>carry on business as usual.  Once there's a mechanism to indicate "hey,
>this document is a flavor that forbids relative URI references", then the
>W3C can move forward on that.
>

The mechanism is in the pipe.  That is, if we can just back out of this
rathole it should be coming out presently.

Is there anything keeping the schema at the end of the schema-location
property from applying such a subtype restriction on the ns-attr?

Al

>Simon St.Laurent
>XML Elements of Style / XML: A Primer, 2nd Ed.
>Building XML Applications
>Inside XML DTDs: Scientific and Technical
>Cookies / Sharing Bandwidth
>http://www.simonstl.com
>

Received on Wednesday, 7 June 2000 13:43:42 UTC