[
Lists Home |
Date Index |
Thread Index
]
At 03:18 PM 4/23/2003 -0400, John Cowan wrote:
>So it can, provided you remember that its presence is merely a hint and
>Schema-based tools need not respect it.
Well that is sort of another problem depending upon what not respect means.
For me it isn't a hint as much as it is a declaration of conformance and a
label. The other issue for me is that there is nothing to say what the
intended environment is/was.
So what I'm really asking for is:
- some sign in the XML that is required to indicate that I'm validating
the document (doctype or an attribute works for me)
- if it is DTD or schema based (for there is only one schema, but it
could indicate others)
- a unique identifier to tell me what the sender modelled their data
stream against
Now what a receiving system/person does with this is up to them. In DTD
land we had catalogs to override the local definition or placement of the
DTD, and map that to my own copy of that DTD. To me this says the system
has read and respect the senders intended environment and what I then
change it to is up to me, but the document is a signed contract telling me
critical pieces of information.
I work for an organization that certifies the XML streams of the members to
conform. If someone sent me a stream and says we process with a DTD only
and I do the same then create. I then use the schema to better check the
content of the message per the data types. But if someone says we have a
schema based process (in this case we only allow W3C schemas) and it fails
to validate - then I've got to question what they are doing if they can't
comply to this first simple step.
>RFC 3151 (http://www.ietf.org/rfc/rfc3151.txt) explains how to map a
>public identifier (either an FPI or not) into a URI suitable for use
>either as a namespace name or as a schema identification hint.
Thanks for the pointer, but the FPI was just an example. I can live with
unique URLS of any form, but I need them to be required to indicate schema
processing.
> > But how many people are working with more than one schema?
>
>For me it's routine to use different schemata at different stages of
>processing.
Ok, I hadn't really considered this situation or mapped it to what we
actually do. We have the W3C schema and the DTD and in the future we might
support others. So can we start some sort of XML Catalog like approach to
managing DTDs and schemas of all forms and specification?
..dan
|
