One issue that has been raised to me by non-TC members is
how the SPML 2.0 schema implements the Open Content Model. The complaint has
been made that the SPML XSDs violate the Unique Particle Attribution (UPA). The
UPA rule basically states that if you want to include any elements, you should
only include the wildcards under elements where there are no other
non-wild-card elements. This allows all the non-wild-card elements to be
validated.
If you want more information about the UPA rule, see: http://en.wikipedia.org/wiki/Unique_Particle_Attribution.
In most cases the Open Content Model is implemented this way
in SPML 2.0. Playing around with this, I discovered that by making a few
changes in how the Open Content Model is implemented, most of the UPA
violations can be eliminated. If we made these changes in the errata, there
would still be UPA violations is Batch, Search, and Bulk (because it uses
Search). All the UPA violations could be eliminated from the Core XSD, which is
the most important one.
I’m not sure that there is much practical value to
this, since in SPML 2.0 all the data is included via the Open Content Model,
but it would reduce (although not eliminate) the places where compilers and
other tools complain about the XSDs.
I added a agenda item to discuss this on Tue’s call. I
want to stress that we can only do this through errata if it would not be
likely to break any existing implementations of SPML, either in the DSML or XSD
Profiles.
|
Jeff Bohren |
|
|
|
13577 Feather Sound Drive, Suite 200, Clearwater, FL
33762 |
|
tel: 727.561.9500 x35719 |
|
|
|
|
