Lists Home |
Date Index |
Rich Salz wrote:
>> XML canonicalisation. In particular, I want to avoid the complexities
>> that result from combining XML 1.0 canonicalisation with namespaces
>> (e.g. Exclusive XML canonicalization :-)
> Once you done the work of C14N, exc-c14n isn't that much more. :)
> Less flippantly, I took a look at RIG2 and skimmed RIG3. It seems
> like the combination of limitations on xmlns (not at the docroot, no
> bleed-thru, no use of alternate prefixes, etc) means that you're
> essentially outlawing the ability to use a layered SOAP
> implementation. Was that the intent?
I'm not sure I follow. I should point out that RIG100 specifies a
business level envelope in which all messages are encapsulated.
Particular transports - such as SOAP - can add their own transport-level
enveloping as required. Thus SOAP enveloping is something that happens
outside of, rather than inside of, RIG 2 compliant messages. In fact, I
expect that most, if not all transport adapters for the SOA as specified
will elect to hive off the RIG 2 compliant payload as an attachment.
This is fine because the transport adapter is the outer most layer of
the SOA onion and is peeled prior to content-based routing,
interventionist proxying etc.
> Also requirement 25 in RIG2 (no superfluous declarations) seems
> redundant since RIG2 already mandates RIG3.
Thanks. I'll look into that.