- From: Rick Jelliffe <rjelliffe@allette.com.au>
- Date: Mon, 09 Mar 2009 19:10:38 +1100
- To: xmlschema-dev@w3.org
- Message-ID: <49B4CEFE.40005@allette.com.au>
The initial beta release of my open source XML Schema validator is
available now, from Schematron.com
<http://www.schematron.com/resource/XSD2SCH.zip>.
I will put out a fresh version with explicit licenses in a couple of
days, pending any immediate bug reports. The release is strictly beta,
prompted by several requests. Improvements are highly welcome!
This implementation takes the approach of converting the XML Schema to
Schematron code using a pipeline of XSLT2 scripts. The methods used have
been explored and documented in this blog series
http://broadcast.oreilly.com/2009/03/post-1.html
The coverage is approximately:
* *simple datatypes:* believed to be 100%
* *list and union datatypes:* not supported
* *structural constraints on elements and attributes:* supported*
* *multiple namespaces, import and include:* supported
* *identity constraints:* not supported
* *dynamic constraints:* (xsi:type, xsi:nill) not supported
* *tricky prefixes:* (elementFormDefault) not supported
* Content model validation is implemented by a series of finer sieves,
which combine
to provide most of the capabilities of a full grammar checker. If a
grammar has repeated
particules or complex nested occurrence constraints, there may be some
false positives where our sieves are not fine enough, however there are
never false negatives.
At some stage in the next few months, a proper open source home will be
found for this project. In the meantime, please report any bug fixes to
me. (I probably don't have time to fix bugs or gaps in any timely
fashion, but of course you can send bug reports too. But fixes will get
everyone further.)
For announcement details and documentation, see
http://broadcast.oreilly.com/2009/03/post-1.html
The current version is a validator developed with sponsorship from
JSTOR, whom I would like to thank.
The advantage of the Schematron schema is that it may be easier to allow
user messages to be customized than with a compiled validator, however
because grammars don't give away anything with respect to the intention
of the schema-creator, the validation messages are still limited to
fairly generic messages, unlike a specific Schematron schema. If ISO
Schematron gets the mooted "properties" features, then it would be
possible to generate a version of the PSVI, with type annotations,
defaults and so on in an SVRL document linking to each element. There
are no plans to do this, as it seem it would be very heavyweight, unless
there is some demand.
Cheers
Rick Jelliffe
Received on Monday, 9 March 2009 08:11:09 UTC