FYI: use of RELAX NG in the RuleML 1.0 Release

[Robin Cover <robin@oasis-open.org>]

I could not immediately locate a post to this list on the use of
RELAX NG in the RuleML 1.0 Release, so see below, if this is
not redundant.

- Robin

[1] RuleML 1.0 Release
     [posted Sat Jan 28 20:06:56 AST 2012]
     http://mail.ruleml.org/pipermail/ruleml-all/2012-January/000325.html

[2] from RNG list (Relax NG, Jing/Trang issues)

Hi Relax NG Colleagues,

RuleML Version 1.0 has recently been released:
http://mail.ruleml.org/pipermail/ruleml-all/2012-January/000325.html

This (fine-grained, freely-combinable modular) RuleML release

* is the first to use Relax NG

* still maintains XSD for 'Rosetta'-style comparison

* is the last planned to use hand-written XSD
  (RuleML 1.1 etc. are planned in Relax NG only, with XSD generated from it)

We hope for your feedback on RuleML 1.0, which should be sent directly
to ruleml-all by February 29, 2012
(http://mail.ruleml.org/mailman/listinfo/ruleml-all).

Feedback on RuleML as a use case for Relax NG should be sent to the
current list.

There are a number of (XSD and Jing/Trang) issues that are causing the
generation of XSD from our modularized Relax NG to be not fully
automated. These include:

1. Inability to reproduce our Relax NG modularization in XSD, thus
requiring translation for each sublanguage.
Due to the large number of possible sublanguages, these
transformations will have to be performed on-demand through a web
application. We use a multi-step process, where the Relax NG is
simplified (either Jing -s or a custom flattening), then converted to
(monolithic) XSD with Trang.

2. Jing/Trang Issue 51
(http://code.google.com/p/jing-trang/issues/detail?id=51&sort=id&colspec=ID%20Type%20Status%20Priority%20Stars%20Module%20Summary)
For the first step of the above transformation, all we really need is
the flattening of the modular Relax NG schema to remove "include"s.
The lack of this Trang option causes us to (currently) use Jing -s
instead, leading to greater modification of the schema than we really
need. For example, we use meaningful names (formula-assert,
formula-query) for context-dependent patterns, which are replaced by
number-indexed names (formula_1, formula_2), making the resulting XSD
hard to read. We are exploring implementing a transformation that
would flatten the Relax NG modularization without otherwise
simplifying. Has this been implemented elsewhere?

3. Jing/Trang Issue 140
(http://code.google.com/p/jing-trang/issues/detail?can=2&start=0&num=100&q=&colspec=ID%20Type%20Status%20Priority%20Stars%20Module%20Summary&groupby=&sort=id&id=140)
also causes trouble because "notAllowed" is used throughout our
modularization. We could avoid this problem if we had the flattening
option mentioned above (either Trang or custom) which would not
perform such simplifications.

We would be interested in any recommendations - please respond on this
list, as these topics are of general interest to the Relax NG user
community.



-- 
Robin Cover
OASIS, Director of Information Services
Editor, Cover Pages and XML Daily Newslink
Email: robin@oasis-open.org
Staff bio: http://www.oasis-open.org/people/staff/robin-cover
Cover Pages: http://xml.coverpages.org/
Newsletter: http://xml.coverpages.org/newsletterArchive.html
Tel: +1 972-296-1783