Duplicate Message Id

Mar 15, 2010 at 4:33 PM

We're raising this issue as a topic for discussion for the Beta. 

For the proof of concept our service will process a HTTP request with a message id that has already been used to simplify testing, however this probably shouldn't be allowed for the Beta and going forward. 

Going forward, in what stage of the workflow should this error be reported?  Should it be reported in the synchronous response, such as 'malformed HTTP parameters' since the message id is invalid?  Or should it be checked later in the processing and be reported in the asynchronous response?