Re: Namespaces and ixml

Steven Pemberton <steven.pemberton@cwi.nl> writes:
> I proposed a way of doing namespaces using the existing mechanisms:
>
>  data: @xmlns:iso, iso:date+.
>         @xmlns:iso: ^"http://example.com/ns/date".
>  iso:date: ...etc...
>
> My feeling is that even if there were another notation added as well
> for namespaces, the method above would have to work anyway. So why
> would we want to have two methods of doing the same thing?

For context, but also because I feel strongly about it, let me repeat
what I said before: that notation cannot be allowed because the “@” mark
(is expliclty described as one that) serializes nonterminals as
attributes and namespace declarations are not attributes.

That proposal is inconsistent with namespaces in XML, would be difficult
for users to understand, and would be unnecessarily burdensome to
implement.

I struggle to understand why there is so much resistence to a proposal
for doing namespaces “correctly” if there is any sympathy for doing them
at all, per the sketch above. If opposition to the proposal was framed
on the grounds “ixml doesn’t need namespaces and shouldn’t have
namespaces at all” I’d be less confused. But I’d also be staring down
the barrel of a long argument about why “@” wasn’t needed either, so
maybe I should be greatful.

It occurs to me that one concern is possibly that adding namespaces to
the ixml specification would be a difficult, would take more time than
we have, and would add significant complexity to the specification.

In an effort to demonstrate that that is not the case, I offer PR #79,
readable here: https://ndw.github.io/ixml/namespaces.html

In summary, as the change markup shows, the full set of changes consists
of:

1. Two new phrases
2. Four new paragraphs.
3. Changes to two grammar fragments and the addition of one new
   fragment.
4. Two new headers, in my editorial style anyway
5. Five new errors.

Questions of editorial style aside, I don’t believe any more is
necessary to implement support for namespaces in a way that is
consistent with modern practice and imposes *absolutely no* burden on
users who choose to simply ignore namespaces and write nonterminals
without colons.

                                        Be seeing you,
                                          norm

P.S. In the event that the proposal to add namespaces to ixml fails to
achieve consensus, the paragraph about nonterminals in the proposal in
PR #79 must still be added to the specification, but it can be
simplified slightly:

  The XML Recommendation reserves names that begin “xml”. Nonterminal
  names should not begin with the letters “xml” in any combination of
  upper- or lower-case. A nonterminal marked as an attribute must not be
  named “xmlns” (error S??).

--
Norm Tovey-Walsh
Saxonica