Schema maintenance procedure

[MURATA Makoto (FAMILY Given)]

Dear colleagues,

In this mail, I propose a conservative approach for maintaining schemas.
Simply put, we use the Assembla respository and merged schemas for
information only.

I have merged transitional XSD schemas and strict XSD schemas.  The
differences are indicated by <strictOnly> and <trantionalOnly>
elements.  It is easy to generate strict or transitional schemas using
simple XST stylesheets.

Every file is available at the Assembla subversion repository.

https://subversion.assembla.com/svn/IS29500/trunk/Parts1And4/XMLSchema

I have also merged trantisional RNG schemas and strict RNG schemas in
the compact syntax.  Having tried quite a few possibilities, I
conclude that we should use the unified format of the diff command for
representing trantisional RNG schemas and strict RNG schemas.  We only
have to specify "-U 10000" and then specify the transitional and
strict RNC files.  It is easy to automatically generate strict or
transitional schemas from the merged schemas.  This approach has one
limitation: when there are no differences, the output is empty.  But I
do not think that this is fatal.

Again, every file is available at the Assembla subversion repository.

https://subversion.assembla.com/svn/IS29500/trunk/Parts1And4/RELAXNG

My initial plan was to update merged schemas only and automatically
generate strict and transitional schemas, and to generate Appendix A
(XSD schemas) and Appendix B (RNG scemas) from schemas.  However, 
in the near future, changes to 29500 schemas are likely to be very small. 
I thus conclude that we should use a conservative approach:

    - We won't update merged schemas, but update transitional schemas
      and strict schemas.  Merged schemas will be generated for
      information.
 
    - The project editor continues to maintain schema appendices of
      29500 as part of DR log chunk files.  Whenever a schema fragment 
      is incorporated in to the DR log, I will check in the change to 
      the assembla subversion and make sure it is syntactically
      correct.

Let's discuss about this in the next teleconf.

Regards,

SC34/WG4 Convenor
MURATA Makoto (FAMILY Given)