Re: [DR 203] Lotus "no" (D) vote on 203

It is unclear to me how we can claim to be providing an RPC if we cannot 
test it.  This seems like an important use case and, hence, an important 
test.  Otherwise, it seems to me that we will wind up with an RPC 
binding that does not serve what I thought was the whole purpose, and we 
may as well be using XML Schema-specified models.

What is the purpose of section 5 bindings if not to provide automatic 
bindings to interoperable representations for existing RPC systems?  How 
can we claim this if we do not test this?  What IS testable, that 
clearly fulfils this RPC purpose?

Ray Whitmer
rayw@netscape.com

> The proposed requirement states:
> 
> "The XML Protocol will guarantee that RPC messages that encode parameters
> and results using the default encoding for the base set of data types will
> be valid for any conformant binding of the RPC conventions. "
> 
> We should specify what goes on the wire, and should ensure that XP is
> suitable for certain purposes.  I don't see how the above proposed req't
> can be meaningfully specified and tested.  First of all, I think the term
> binding here is used to mean binding to programing langs. and object
> systems, which is an inconsistent use of the term wrt the rest of the
> specification.  More fundamentally, I think the requirement specifies
> characteristics of particular bindings, which are beyond the scope of the
> spec.  No matter how good XP is, I can always build a faulty language
> binding for it.
> 

Received on Monday, 13 November 2000 22:21:49 UTC