Re: New draft published

[John Cowan <jcowan@reutershealth.com>]

Paul Grosso wrote:


> But in my view, the catalog isn't returning an system id *to the XML processor*,
> it is part of an entity manager.  That entity manager may well be able to handle
> a fragment id or a ?xxx query string or whatever, and the entity manager then
> returns stuff to be parsed to the XML processor which is none the wiser that there
> was a frag id involved at some point.  Specifically, I want to be able to have the
> rhs's be queries into an XML data base, for example.


You are confounding two cases.  A query string (with leading "?") *is*
part of the URI, and is server-side.  A fragment id (with leading "#")
refers to a portion of an entity body, and is inherently client-side.
We (i.e. the Core WG) decided not to burden XML parsers with fragment-id
processing, which might involve XPointers.

Queries into a database would surely involve "?", so they are all right.

-- 
There is / one art             || John Cowan <jcowan@reutershealth.com>
no more / no less              || http://www.reutershealth.com
to do / all things             || http://www.ccil.org/~cowan
with art- / lessness           \\ -- Piet Hein