Re: PIs considered harmful Was: XML-SW, a thought experiment

IMHO

From: "Elliotte Rusty Harold" <elharo@metalab.unc.edu>
[...]
> Perhaps schema languages should be written
> in a more permissive fashion so that they automatically allow
> anything from other namespaces.

Or maybe there should be a half-way point.  There should be
a standard element type (which any language can subtype) whose
elements in a document have an FYI status. That is,
their understanding is optional and they carry their own significance
independently of the rest of the document, rather than
modify the  effects of other things in the document.

Then
- anyone could, in a new namespace, create new FYI elements, and
- any schema could indicate that FYI elements were allowed in most/all
places.

Also, anyone processing such a document and finding an
unknown element would know that it must be an FYI element
can can therefore ignore it.  So the effect of PIs is reintroduced,
but using the ability to parse and disptach off them and the
other things which already work with elements.

I understand Norm's need, and respect various people's protestations
that PIs are useful and needed now. If support of them continues
light, then I guess leaving them in seems reasonable.  If support
of them starts to get heavier, then it would be wise for the future
to move to using FYI elements, rather than rebuild element support etc.

Of course the otehr tack is to make PI syntax the same as that of elements
with the namespace support so that the complixity would be reduced by using
exactly the
same machinery, and basically the "?" beoming an "optional" flag.

(more random thoughts ... IMHO this does not  consitute
"institutional" commitment!)

Tim

Received on Friday, 22 February 2002 10:57:37 UTC