Re: draft-dolson-plus-middlebox-benefits (was RE: Review of draft-mm-wg-

On 19/04/17 01:33, Martin Thomson wrote:
> On 11 April 2017 at 18:51, Stephen Farrell 
> <stephen(_dot_)farrell(_at_)cs(_dot_)tcd(_dot_)ie> wrote:
> > (*) That is the argument for draft-mm-effect-encrypt, for which I
> > do support publication (apparently in disagreement with Martin in
> > that case:-)
>
> Taking a break from email is so refreshing...
>
> I would very much like to see a document like this that captures these
> use cases and practices.  The problem I have with it are simple: it
> isn't particularly coherent, 
Hmm. Two things to consider there: 1) the editors have AIUI tried
hard to collate offered text from various folks and that's fairly
likely to end up less coherent, so we shouldn't insist on a level
of perfection here that we'd not ask from something with which we
might more "natively" agree.

But possibly more interestingly: 2) I think it's arguable that
coherently arranging the set of descriptions here is hard, and I
think that may tell us something about the underlying logic of
the situation.

For me the lesson of that innate incoherence is that the set of what
are now shown (in 20:20 hindsight) to be somewhat "lazy" practices
are being demonstrated as problematic via the re-assertion of the
e2e argument using confidentiality mechanisms, even where those latter
confidentiality mechanisms aren't in fact deployed "fully" e2e. And
that argument is quite close to the ossification one that transport
folks seem to like. The incoherence of the argument as to why
confidentiality is "problematic" seems to me more a fundamental
weakness in that argument against confidentiality rather than a
failure of exposition in this particular draft.

> and it makes statements that I disagree
> with.  Those statements are particularly bad because they are also
> unnecessary 
I do agree that the draft ought not be recommending things or to
be saying that such-and-such a practice is "necessary." But I did
think we'd gotten rid of those. Mopping up remaining statements of
that sort would be good, I agree. There were a lot of them in the
contributed text that the editors already excised. (In saying that
I did not go through your detailed review checking stuff - now that
I've escaped the IESG I'm fine that someone else does that:-) And
of course, the fact that you or I only "disagree" isn't really a
good criticism, we need an argument that those statements aren't
well grounded or similar, but I guess that's what you meant.

Cheers,
S.

> (see Brian Carpenter's comments about facts).

signature.asc
Description: OpenPGP digital signature