Re: Issue 169: Propose http method in the operation interface to simplify http binding.

Hugo Haas <hugo@w3.org> writes:
> 
> To answer Sanjiva's question, I agree that Paul that the relationship
> between @webMethod and @wsoap:mep depends on the binding. If the
> binding supports the SOAP Web Method Features, these will be hints for
> how to bind operations.

Sounds reasonable to me.

> In the case of the HTTP binding, because of [3], I believe that we
                     ^^^^^^^^^^^^

You meant to say the SOAP/HTTP binding as defined by XMLP and as
used by us in the WSDL->SOAP binding (wsoap:*), right?

> need to say that:
> - @webMethod=="GET" should bind to SOAP Response as @wsoap:mep (if the
>   URI style is in use, otherwise it wouldn't be possible);
> - @webMethod=="POST" must bind to the SOAP Request-Response MEP as
>   @wsoap:mep;

These two rules would make @webMethod redundant given we already
have @wsoap:mep. We say now (following the rules of XMLP AFAIK) that
if @wsoap:mep=SOAP-R-R then the HTTP method must be POST and that
if @wsoap:mep=SOAP-Response then the HTTP method must be GET.

> - @webMethod=="PUT" and @webMethod=="DELETE" must bind to the SOAP
>   Request-Response MEP too;

For those values the XMLP-defined SOAP/HTTP binding does not apply,
right? If so we have nothing to say about it .. at least as our
spec is today.

> - for other values, no recommendation is done.

What exactly is the proposed value space of @webMethod?

Sanjiva.

Received on Tuesday, 13 July 2004 02:19:09 UTC