Re: Issues remaining with Bind draft

Lisa wrote on 03/17/2004 06:08:26 PM:

> General
> -----------
> 
> The spec must stand alone, not be dependent on changes to RFC2518 in 
> 'bis'.  Otherwise, bind can't be approved until RFC2518bis is approved. 

I wouldn't describe this as "standing alone" (since it clearly must
depend on RFC2616, RFC2518, etc.), but I agree that it cannot depend
on material in specs that have not yet been finished/approved.

>   That means no dependencies for things like 'lockroot'.

As Julian indicated, we are not aware of any such dependencies.

> In general, the spec needs more info to specify how existing things 
> work. 

The spec needs to describe how the new things it is
defining work, and it needs to describe any modifications to the
behavior of existing things.  But if it is layered above some other
specs (as the binding spec is layered above 2616, 2518, etc), it
should not repeat or reword behavior it is just inheriting unmodified
from those other spec.  Doing so would only introduce confusion
when the spec that does define the behavior of those existing things is 
revised.

> All the following questions must be answered in the spec, NOT 
> just in email.

If every question that ever arises about a spec on the
mailing list resulted in a paragraph or even just a
sentence being added to the spec, the spec would rapidly
become too bulky and verbose to be useful as a spec.

> The spec must be explicit, because different people 
> reading a model description always end up with different ideas how the 
> model works in practice.

As Julian indicated, there are a variety of sources of information
that are used to help implementors.  The spec is of course a critical
piece, but the issues list and the mailing list are two other essential
pieces.

So every question gets answered in the mailing list.
Some questions get added to the issues list.
Some issues result in changes to the specification.

Cheers,
Geoff

Received on Thursday, 18 March 2004 09:02:20 UTC