Are PI targets arbitrary?

james anderson James.Anderson at
Mon Jan 25 15:48:04 GMT 1999

One way in which this differs from namespaces is that it does not provide a
mechanism to avoid name conflicts. A consistently applied encoding mechanism
for namespaces would have. That is, the ns-spec should well have premitted all
names to be qualified names. Since it precludes the name in a notation from
being qualified, it is not possible to use notation to this end. I have heard
the argument, that the association with a notation and the uri specification
for a notation obviates the need for qualification. It didn't understand the
claim. I still don't.

One really needs to an encoding of the form
<?pi fop:some-command some args ?>

which together with the binding < ... xmlns:fop=""
... >
would permit the application to operate with a token of the form
 ("" X "some-command")
and be assured that there are no name conflicts.

Yes, the present spec would permit this, but does not describe its effect.

James Tauber wrote:
> I'm starting to put some options into FOP that are triggered by processing
> instructions. My initial thought was use 'fop' as the target.
> But then it occurred to me: PI targets can be associated with a URI via a
> NOTATION declaration. This is *a lot* like namespaces. Does that mean PI
> targets should be arbitrary proxies?
> Instead of looking for PIs with the target "fop" should I instead be looking
> for PIs with the target which is a NOTATION declared with URI
> After all, some other application might use the target "fop" but only mine
> would use the NOTATION with URI

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list