spf-discuss
[Top] [All Lists]

RE: AOL Spam down 27%

2004-03-25 15:56:01
From: wayne
Sent: Thursday, March 25, 2004 3:52 PM


In <MHEGIFHMACFNNIMMBACAOEIIHKAA(_dot_)sethg(_at_)GoodmanAssociates(_dot_)com>
"Seth Goodman" <sethg(_at_)GoodmanAssociates(_dot_)com> writes:


<...>

 "Good" ideas are a dime-a-dozen.

Ideas are a dime-a-dozen.  Good ones are few and far between.


"Good" ideas without an
implementation often don't look so good after all the subtle problems
have come out when you try to put put the idea into action.

Because implementations are expensive and talk is cheap, it makes a lot of
sense to first discuss an idea before running off and implementing it.


"Good"
ideas without implementations aren't particularly useful for most
people.

This is completely contrary to my own experience.  In my work, I usually
work in an interdisciplinary team.  I propose architectures for transducers,
circuit designs, communications channels and sometimes software.  Other
members of the team critique the architecture and suggest alternatives.
Though those team members couldn't create an implementation of what they
suggested, their ideas are often very useful.  If I insisted that they prove
their ideas worthy of consideration by creating an implementation, the
design review would be pointless.  More importantly, the resulting
architecture would be poorer for the lack of alternative ideas.  In fact,
some of the better ideas came from people who knew the least about circuit
design but were still smart enough to evaluate an overall architecture,
simply because they tend to look at the problem from a different
perspective.  Everyone on the team has advanced degrees, decades of
experience and is secure enough in their own knowledge that they can listen
to ideas from non-specialists.



I should note that the idea of a designated sender has been around for
many years.  I implemented a very simple one over a year ago and added
a plugin to SpamAssassin to support it.  At the time, none of the
other designated sender proposals had working code.  When Meng put up
the SPF website, he also put up working code.  That is the primary
reason I abandoned my implementation and started pitching in here.

Action speaks louder than words.

The proposals that I posted on SRS-discuss are not working code, but they
are quite a bit more than just random ideas.  As you will recall, I also
posted a couple of early but more detailed versions to a subset of this
group.  IMHO, that amount of effort constitutes action.

Yes, designated sender is not a new idea.  However, combining it with SPF in
the particular way proposed appears to cause a lot less breakage and results
in more benefits.  The benefits are substantial.  I think we would all
benefit if you would consider and critique those ideas.

--

Seth Goodman


<Prev in Thread] Current Thread [Next in Thread>