xml-dev - Re: XML Schemas: Best Practices

[Francis Norton <francis@redrice.com>]

"Roger L. Costello" wrote:
> 
> What do you think?  When you create a schema component should that
> component be expected to have the same semantics regardless of the
> application that uses it, i.e., universal semantics?  Or, should the
> component be able to "semantic-morph" to each application, i.e.,
> localized semantics?
> 
Syntax has to serve a purpose to succeed. But the "purpose" might be as
specific as a UK 2001 tax return or as general as XML-Schema. (How do
you tell the difference? Which of the two has a clear, black-and-white
functional requirement and a single, simple test for its success or
failure?)

Single-purpose grammars will have different change and evolution issues
from multi-purpose grammars. If you borrow a component from a
single-purpose grammar for different purpose you may find that the
grammar and associated tools evolves in ways that make it unsuitable for
your requirement.

I think that successful grammars will have clear documentation about how
specific or general their semantics are. Best practice in designing and
using schema components will be to follow best practice for component
design and usage in general, including allowing for evolution, and to
document how specific or flexible the component is intended to be.

A last thought - sometimes looking at how to do something wrong can cast
light on how to do it right. I am told that one of the problems of EDI
is that the spec is so rigid, and it is so hard to achieve official
variation (top-down authorisation required) that normal working practice
is to grab unused elements and re-use them to accomodate elements
missing from the official schema - I've certainly seen this happen in
corporate databases. How does that effect XML Schema best practices?

Francis.
-- 
Francis Norton.

why not?