[dsdl-discuss] Re: Potential use cases for complex value validation from ebXML

From: "Martin Bryan" <mtbryan@sgml.u-net.com>

> Unless we ensure that validation rules are
> clearly dated we cannot be sure that we can revalidate a document at any
> time after its initial validation, because someone could have updated the
> external code list between validations.

Could we say there are three kinds of validation for controlled vocabularies?:

* feasible -- the lexical rules for names of that kind of code are followed
* possible -- the code has been used in some version of the controlled vocabulary
* specific -- the code is used in a specific version of the controlled vocabulary

"Feasible" (which is similar to WSX's "lexical space") is the minimum that people
need for validation, but it is not sufficient for UI.

"Possible" allows all documents using a vocabulary of any version. Good for validation and OK for UI. But it must be maintained.

"Specific" probably requires some level of co-occurrence constraint,
where the document instance states which specific version of the vocabulary is
used. This requires
 1) a repository and/or
 2) somewhere to park multiple controlled vocabularies inside the DSDL file

The latter is what the "schema" section of that Schemachine strawman allows.

An issue related to this is how to handle when Web resources are unavailable.
WXS has an approach on this. A web resource may become unavailable because
the LAN is down or because the resource has been deleted.

I am very suspicious of validation relying on third-party resources. It seems
to mix layers incorrectly. If the schema is to be found in a third-part repository,
then there is some other layer that should take care of fetching it and handling
exceptions before we get to DSDL. Not clear on this one...

Cheers
Rick Jelliffe

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)