Hi, Sanjiva --
StAX parsing won't permit schema validation, so that's a no-go
there. And schema validation isn't truly enough, as there are valid
but illegal BPEL processes.
Um why is validation a mandatory thing? If you want to validate before
reading then you can do it using various forms- IMO forcing validation
upon reading is not necessary.
Validation is a desirable thing, at least IMHO. Is there a QName
where there is supposed to be a QName? Is the content of an
expression in the right place? Is the namespace of the root element
correct? Ultimately, validating against the schema and building the
toolchain to only accept valid (and semantically correct) BPEL is a
contract with the user -- we'll do what you want/expect if you supply
correct input. (Semantically-correct BPEL is a subset of schema-
valid BPEL.)
I have previously done the work of translating the XML Schemas for
versions of BPEL into RELAX NG, and there is a good amount of
meaningful information (e.g., exclusion of the presence of one
attribute based on the presence of another) that the XML Schemas do
not include. All of this is meaningful for BPEL -- If both
attributes are present, which do you want the engine to use? The
schemas or grammars are a kind of program that suits the problem
domain and are generally agreed to meet the need.
Of course, it would be nice if people would just define real grammars
if they're going to make programming languages, but I digress... :)
The other aspect of the object model design must of course be that one
can create instances of the model in memory and run with just that
instead of having a .bpel XML file around at all.
The BOM is intended to be a developer-friendly (the things have the
same names in the OM as they do in the BPEL) object model for BPEL --
add activities to a scope, etc. It was ideally conceived to be both
the target of the parser component and a potential model for
something like a graphical designer or MDA toolchain that had nothing
to do with XML at all.
--
Paul R Brown
[EMAIL PROTECTED]
http://mult.ifario.us/
