Robert A. Morris
At TDWG, David told me the argument against using SOAP in DigIR, but it was so unconvincing that I
can not reproduce it one week later.
It seemed to hinge on the requirement that SOAP messages be transported with http POST but some argument I can not reproduce that GET is needed. Since I would assume that queries are in the payload, not the headers, I can not understand why there is any problem passing URL encoded queries in the payload. Further, if a data provider has to implement /some/ protocol, there is no additional development effort required on providers to implement SOAP, which, after all will already have lots of infrastructure support that would have to be rebuilt if the project invents a new messaging protocol.
Does anyone care to elaborate on the opposition to SOAP?
I'm with you Bob.
SOAP seems to have wide acceptance and the tools/infrastructure to match. I too would like to know the weakness in SOAP as the transport mechanism for a DiGIR package.