Re: Fixed base

>My point was (and is) that there is an association between
>the namesapce URI and a set of elements -- e.g. the HTML element
>set, or the Dublin Core -- with their associated rules for content
>and, where appropriate, behaviour.

Granted. But those assocations are being asserted by the content/behavior
descriptions, rather than by the Namespace... not least because there may
be several such metadata sets which reference the same namespaces for
different (though presumably related) purposes.

XML Schemas certainly indicate which namespaces they expect to see.
Namespaces do not, at this time, inherently state which schemas they belong
to.


>>> and also as a hook to activate specific behaviour.
>> No more so -- and no less so -- than the localnames.
>Oh.  There are implementations that attatch semantics to the
>use of a specific prefix?

Please reread: I said "localname", not prefix or qname. Localname and
namespace   are co-equal parts of the expanded name of the element or
attribute -- as defined by the namespace spec -- and the combination of
these is used to key semantics.

The prefix, we all agree, is syntactic sugar and should play no direct role
in this. (The decision in some specs to refer to prefixes inside textual
values is rather unfortunate, in this regard, though I understand why
people thought it was a good idea at the time.)

Received on Thursday, 22 June 2000 12:37:34 UTC