Paul,

>> When folks agree on the new mail transfer services that we need and
>> when we try to add them to smtp and fail, THEN we can have productive
>> discussions about a replacement transfer protocol.
PV> well, except that that's not how dns was created, or http, or html, or
PV> nntp, or xml, or rpc/xdr/nfs, or sip, or pgp, or jabber.

_New_ services get created in all sorts of ways and for all sorts of
reason.  However changing an existing, popular service is subject to
very different concerns than a new service.  In particular, it is
subject to careful attention to preservation of the installed base.

Facile assumptions that we will blithely move an installed base of 500
million people, to a new set of protocols, reminds me of a cliche
cartoon. It has a blackboard -- it's a very old cliche -- almost
filled with formulas, except for one blank strip that separates the
formulas into two sections. The board is being explained by one guy to
another. He is responding to a question, pointing to the blank space
and says "that? oh, that's where a miracle happens."

Switching 1/2 billion people requires quite a lot of force and time,
and so do the hundreds of thousands of implementors and operators who
have to make it happen.  They need clear and compelling incentives for
the considerable energy it will take and discomfort it will cause.

So far, claims that smtp needs to be replaced, to fix spam problems,
fail to provide anything more compelling than some strong emotions.

Let's remember that no action to date has reduced the global level of
spam.  So folks need to be a tad circumspect when calling for massive
infrastructure change for which there is no basis to guarantee
results.

>> And everyone else needs to move from the generic reference to
>> "consent" on to something that is more concrete, as well as being
>> integrated into a full range of human uses for email.

PV> i'm pretty comfortable with www.dictionary.com's definition of "consent".

Me too, for casual discussion, but it has nothing to do with technical
specification nor for careful understanding of the human and social
dynamics of messaging.

Really, Paul.  Pursuit of spam control needs far more detail and
deliberation.  It's urgent, but that's no excuse for being vague and
generic.

At a minimum, claims that we need to replace smtp need to include a
specific proposal that offers specific features absent from smtp. And
it needs to include a transition plan for those hundreds of thousands
of operators and 1/2 billion users.

d/
--
 Dave Crocker <dcrocker-at-brandenburg-dot-com>
 Brandenburg InternetWorking <www.brandenburg.com>
 Sunnyvale, CA  USA <tel:+1.408.246.8253>


Reply via email to