[
Lists Home |
Date Index |
Thread Index
]
That requirement doesn't make sense AFAICT:
1. As you say, the customer has the DTD. Why
do they want to invoke a network process to a
web machine to validate a local document?
That is not only more work, it is error-prone
and opens up a security hole if you do something
untoward.
2. Now they have to negotiate with you anytime
they want to change their own DTD. They are
as Bill Gates puts it so well, 'outsourcing
their brains'.
3. Why are they validating an RPC command?
I understand a requirement that you validate
before sending but not one where they validate
against their DTD copied to your machine.
Without understanding the basis for the requirement,
it's hard to say if the best design pattern is to
give them what they want or to chide them into a
more sensible solution. They don't seem to understand
why DTDs were made optional in XML.
len
-----Original Message-----
From: john Mani [mailto:john@sixthdimension.com]
We have an application that exchanges XML with a customer's
application, running at their site. The XML DTD is specified
by the customer. These XMLs are used more as a RPC type
command, rather than as documents that may be archived and
reused later.
The customer's XML processor demands that the XML documents
we send have a <DOCTYPE .. > declaration, that points to a URL
hosted by us.
I would prefer that I do NOT include any <DOCTYPE > declarations,
but just have the XML's be declared 'Standalone'. The receiver's
parser afterall has the DTD to insure the validity of the document.
And it certainly doesn't make much sense for me to be hosting
the DTD URL, rather than the DTD owner (me customer, in this case).
Anyways, what's a design pattern in such situations ?
|
