Re: Moving on (was Re: URIs quack like a duck)

> You failed to mark the fact that these quotes are from different messages
> at different times, no?

I thought it was rather obvious I didn't mean to deceive, or
criticise... Just a light hearted dig.

> What transaction requires that there be a name?

Every aspect of namespace processing.

> I thought that we had established that there is no need to refer to the
> namespace.

You have to know what the namespace and local name of an element is
before you can process it by any namespace aware tool (of which
xpath/xslt is currently the most well known example, but there is
nothing xpath specific in any of this, assuming xpath and namespaces
are brought into line)

> What depends on the definition?  Actual operations, value added?

I don't understand this question.  _everything_ depends on the
definition of the namespace name. If you don't know the namespace name
then you can't do anything.

> Processors of a particular namespace can always recognize their proper
> instances to process by the matching method provided in the existing
> Recommendation.

That is the literal option. The point is that some people want to
change that, in the case of relative URI, the suggestion was to change
it to be undefined.

> The cited resource could include
> format-sensitive information like a DTD by reference, so that the tool
> machine-applying the DTD wouldn't have to sift through a flabby document to
> find the material germane to it.  Or it could be a "literate programming"
> doucument where the extraction to the/each machinable subset is well
> defined and easily done.

Yes the packaging proposals (which I don't think ever got anywhere)
would amount to recommending a particular format for such a file of
references, so that you didn't end up downloading lots of useful
information in some wonderful literate program that unfortunately you
can't read.

David

Received on Friday, 2 June 2000 13:20:13 UTC