Re: Unserializable documents

On 9/5/07, Henry S. Thompson <ht@inf.ed.ac.uk> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Alex Milowski writes:
>
> > In fact, I'd like to go through all our steps and guarantee that we do the
> > right thing with namespace declarations.
>
> I think that will be a substantial, and unnecessary, undertaking.  My
> _very_ quick survey suggested _at least_ 8 steps which might need
> fixup, and I don't think the same text will do in each case.
> Furthermore, it is _quite_ expensive and tricky, and I'd _much_ rather
> leave it until the margins, i.e. at actual serialization time.

It is hardly quite expensive.  The cost is linear in the number of
elements & attributes
in the document.

In reality, you don't need to do much if you can use an XSLT 1.0 processor
that guarantees namespaces are declared for elements and attributes.  XSLT 2.0
and XQuery already do this and I've outline a proposal that makes the same
guarantee for the rest of our components.  If that is the case, then there is
zero cost because there are no namespaces to fixup.

So, let me clarify this.  My proposal is that all our steps are
well-behaved given
that they receive a serializable infoset (e.g. one that doesn't
require namespace
fixup). All their outputs also do not require namespace fixup.

In all of this, it is just XSLT 1.0 that may cause a problem.  Implementors can
just pick an implementation that guarantees namespaces are declared.  ...or, if
they have their own, they can fix code so that it does so.

-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

Received on Thursday, 6 September 2007 01:09:41 UTC