[e2e] Discrete IP - retake
Dave Crocker
dhc2 at dcrocker.net
Tue Sep 18 09:43:35 PDT 2012
On 9/18/2012 9:17 AM, Jon Crowcroft wrote:
> this is what we used to talk about as the
> "my problem is too hard even for you" poser syndrome
>
> basically, whenever you offer a workable solution,
> the poser (of the problem) changes the
> problem (or the assumptions)
This is one of the reasons that I am a fan of starting any technical
effort by agreeing on a simple, brief, entirely non-technical
description of the problem to be solved and the benefit to be obtained.
In the case of revision efforts, it's especially helpful in providing a
basis for analyzing deficiencies of the existing work, as well as
possible improvements of proposed work. It also forces a focus on
community need in terms of user-level functionality that is not
distracted by technical nits.
Predictably, the specific example I happen to encounter most often is
assertions that we need to replace SMTP with something that has better
security (to stop spam, phishing, and other unpleasant content.)
My suggestion is first to get the community to agree on a non-technical
statement of the services that we need that we don't currently have.
This is, and should be, quite difficult. Changing a complex human
communication system is certain to have very serious and damaging
unintended consequences.
In any event, until that agreement is clear, worrying about the
technical work is wasted effort.
Once the agreement is in place, give the SMTP technical folk a chance to
find a way to implement it. That's a realm of activity that has revised
itself successfully for 30 years, so there's a good basis for thinking
it might be able to incorporate the 'next' set of functional changes.
At some point a revision effort will fail and indeed we'll have to
replace SMTP, but until we go through the sequence, replacement is not
appropriate. The switching costs for an infrastructure service are too
massive.
I of course intend this as an exemplar for all proposals of major change.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
More information about the end2end-interest
mailing list