deprecating BCP38 and similar

I think I'll change my position on BCP38. It's pointless to try
blocking spoofed source addresses because:

* It doesn't solve every single problem
* It means more effort for service providers
* It requires more CPU processing power
* Using it will generate smarter "black hats".

I also think everyone should drop all forms of IP ACLs and
password checking. Neither of those have solved every Internet
problem, they require more effort and CPU, and smarter crackers
have surfaced as a result of their deployment. These measures
are ineffective, and it is silly to waste time with them.

Anyone from Microsoft listening? I suggest you terminate your
Trustworthy Computing Initiative. Not every problem is caused by
a buffer overrun or race condition, and you're wasting billions
of dollars. I suggest you post regularly to NANOG, helping
educate the masses that anything less than a silver bullet is

Eddy, who hopes everyone recognizes hyperbole and sarcasm

with everything you should look at the effort, the returns and the risks. some
simple things can have major benefits but we shouldnt waste effort on major
changes that have little effect and that can be circumvented (i'm referring to
the port 25 blocking discussion of course, wrt bcp38 i dont think anyone [with
clue] thinks its not worthwhile)