RE: Comments on bind-08

Julian writes:

> Are we supposed to put examples that having nothing whatsover 
> to do with BIND per se into the BIND spec just because it 
> happens to be the next spec that's supposed to get finished? 
> I fear that is an open-ended set of possible clarifications, 
> and doing all of them certainly is not the right way to 
> finish the spec.

I think there should be examples of If header use in the BIND specification,
and also in RFC 2518bis. Though quite related, this repetition is
worthwhile. It is worthwhile in 2518bis because of the observed problems
with client implementations of the If header. Interoperability errors are
expensive to identify and fix, and tend to have negative impacts on users
before they are fixed. Examples often help reduce interoperability errors,
and are less costly to produce than it is to identify and fix an
interoperability problem.

The bind specification introduces some new wrinkles for If header passing
and processing, especially the existence of loopback bindings. Additionally,
a diligent, careful specification reader could still find it useful to have
an example that confirms their understanding of how to apply the If header
processing described in RFC 2518 to collections using bindings. For these
two reasons it also makes sense to have an example of If header passing in
the bind specification.

- Jim

Received on Thursday, 2 December 2004 17:06:10 UTC