Re: [apps-discuss] Status of the malformed mail BCP

[Andrew Sullivan <[e-mail]>]

On Wed, Aug 10, 2011 at 09:10:38AM -0700, Murray S. Kucherawy wrote:

> So I think my next step is to convert the existing document into
> exactly that.  Since the main objection to doing it as an RFC is the
> idea that it might need to be changed often, perhaps we should go
> with Expert Review instead of one of the rules that says a
> specification has to exist.  Any comments on that point?

I think this is stretching past the point of pragmatism.

The proposal, so far, is a sort of extreme case of being liberal in
what one accepts (and what one does thereafter, of course).  I have to
agree with Peter's detection of slight absurdity here, but I also
question whether it would be possible to select an appropriate expert.
Who is more expert in not conforming to the existing protocol
documents than the person who is so nonconforming?  And if there is
nobody, then the correct expert to appoint in each case would be the
implementer of the nonconforming implementation, who would perform the
review and approve the registry entry.  Therefore, assigning the
registry entry on detection of nonconformance would be correct, and no
expert is needed. 

At that point, of course, the problem becomes that you need a stable
reference for the way the nonconformance works, and so you need a
specification anyway.  So "specification required" is correct, I
think.

(The entire passage above, however, could as easily be construed as a
reductio argument against such a registry at all.)

A

-- 
Andrew Sullivan
[e-mail]