> I understand that "x-" parameters are effectively defined as
> "known to be an SRU parameter". These are not "unknown parameters".
Not quite - the x- prefix is really a way of "namespacing" the parameter
outside of the SRW/SRU parameter namespace, so that we don't hit any
confusion later if we introduce in SRW 1.x a parameter with the same
name as someone is using as a local "unknown" parameter.
In addition for SRU there are some semantics about how parameters get
echoed back in a manner consistant with the SRW
extraRequest/ResponseData mechanism.
> At issue, I think, is what an SRU server is _required_ to do
> on receipt of a parameter that is not known to be an SRU
> parameter, e.g. "linkspeed=high&" or "offensive=block".
>
> My preference is that SRU would not require servers to fail
> the operation and return a diagnostic whenever a query string
> includes an unknown parameter. Some SRU servers may have that
> behavior, but other compliant servers should be able to
> process or ignore as appropriate to their application.
>
> Eliot
>
>
|