[
Lists Home |
Date Index |
Thread Index
]
Simon,
Yes, I think that there is a class of applications for which the PSVI is
where the action is. This is especially true of complex data including some
data that is strongly typed, and especially true when typed data is stored
in a persistent form. Some of this persistent data may never be physically
instantiated as XML.
The "XML as bytes" and "XML as PSVI" camps will both exist for a long time,
and it's important that the two groups be able to enjoy a red wine together
in Barcelona. I applaud your initiative to say that both groups have a
right to exist.
Serialization is important for the PSVI people too. Data, including the
PSVI, needs to be exchanged. Unfortunately, the current PSVI, though it
contains all the right data, is a bit of a mess. I would like to see a lot
of work put into making it useable and redesigning it. The XML Schema
Formal Description (http://www.w3.org/TR/xmlschema-formal/) is a step in
this direction, but I think there is a lot of work to do here.
I would like to see a "post-processed" XML representation of the
information found in a schema, a representation that is easy to query using
an XML language like XQuery. I would like to see a standard API that is
used in all the language bindings that need access to the PSVI. I would
like to see mappings into this representation from other schema languages.
The XML Instances need not be serialized with the entire PSVI, though there
may be use cases for this. In general, the PSViIcan be computed when the
instance data is received.
None of this should be imposed on the users of XML who just need text.
Jonathan
|
