constraining XRD (YADIS 0.9 section 7)

[Martin Atkins]

Larry Drebes wrote:
> In section 7.3 it would be helpful to differentiate between "YADIS
> constraints on the schemas", and where we are re-stating the schemas as
> they stand.   Two constraints worth removing:
> 
> 7.3.3, "MUST contain one or more Service elements."
> This is not a necessary constraint. True, a document with no Service
> elements is of limited use to an application, but it needn't be an
> invalid case. An example may be someone with a hosted YADIS service with
> no services yet configured. An application will probably treat this much
> the same as a document that contained Services only of types it was not
> interested in.
> Instead of focusing on the minimum requirement here, it might be more
> productive to say what may be included. e.g. "may contain as many
> Service elements as necessary to describe the YADIS services."
> 
> 7.3.4 "MUST contain at least one Type element"
> There is a need for people to know "How to define a new YADIS Service."
> In answering that, we can state that YADIS Service designers should
> define a Type value and require its presence in the Service element for
> use with their protocol.
> "Type" is, however, not a required element in the XRD schema. I do not
> believe a YADIS parser should treat a Service element with no Type as an
> error, as they might do after reading the MUST here.
> 

Agreed on both counts.