correction re: tag prevention bandwidth

[Zooko <zooko@zooko.com>]

Folks:

In my previous message entitled "tag prevention again", I incorrectly stated 
that the two-payloads approach has roughy the same bandwidth consumption as 
the two-headers approach.  This is only true when the real payload is small 
enough to fit into a single message.  (E.g., around 16 KB according to the 
latest parameters I saw mooted.)

If the real payload exceeds that, then the two-payloads approach does consume 
twice as much bandwidth as the two-header approach, although that might not be 
true in practice due to two-payloads being more reliable == more efficient at 
retrying.

Sorry for the confusion!

Regards,

Zooko

Zooko.Com -- Security and Distributed Systems Engineering