Lists Home |
Date Index |
Didier PH Martin wrote:
> Didier replies:
> Tim if I understand you well, your point is that we update the URI specs (or
> create a new URI) in order to be able to perform method invocation. Since I
> do not have access to all TAG discussions can I ask you if the first thing
> that you suggest being standardized is about verb or method invocations?
You do have access to all meaningful TAG discussions:
No! I don't want a new part of the URI space. I want SOAP operations
which have a moderate number of scalar parameters, RPC-like semantics,
and are safe/idempotent, to be available through an ordinary
boring http:... type URL. I've proposed how to do this.
> There are some issues that need to cleared. Issues with actual URIs
> a) size limitation
I don't believe this is a problem any more.
> b) syntax encoding (in some URIs the spaces and other characters have to be
> replaced by a different encoding)
The XML spec gives a recipe for how to do this. Solved.
> c) verb vs. parameters encoding ( amore formal and standard way to encode
> verbs and parameters)
Agreed. See my proposal.
> d) type enforcement when necessary (there are time when the parameters have
> to be conformant to a specific type, the receiving end needs to know that
> and have access to a specification file a la idl, wsdl and tutti quanti).
I'm not trying to cover this part of the problem. I'm trying to hit
an 80/20 point. If you need this stuff, you need the full SOAP
apparatus and you need to do POST. -Tim