Mods to Fuz2 seem to make it less effective

Vernon Schryver
Fri Mar 7 20:41:07 UTC 2003

> From: "Earl A. Killian" <>

> It might be interesting if MANY wasn't the same as a count of
> INFINITY, but instead was a separate count.  E.g. you might have a
> count of 10,3 which means 10 sites saw it without reporting it bulky,
> and 3 sites reported with the bulky attribute.
> It would have to be up to the user of course to decide what to do with
> this extra information if it existed.  Someone that didn't trust
> others' bulkiness attributes might just use the total count, A+B=13,
> whereas others might crunch as before B != 0 ? INFINITY : A.
> This is not a request, just an observation that might make it easier
> to think about it.

My short answer to that is "if you want Pyzor, then you want the office
down the hall."

The longer answer is that without some what to prevent someone from
reporting a single message 17,000,000 times, there can be no difference
between the various infinities.  I believe that one must design to
fit reality instead of the way one wishes things were.  I think it is
impossible to build a system that involves more than a very few dozen
people without including some duds who will do things like reporting
a CERT advisory 17,000,000 times.  That implies that you can only take
(or make) "MANY" or "INFINITY" for what it is and no more.

Of course no one but a bad guy or bad or buggy software that doesn't
detect mail loops would report a CERT advisory 17,000,000 times, but
people do odd things.  For example the logs of the public DCC servers
have plenty of complaints from what I strongly suspect is a pervert
DCC client.  I think someone ported the DCC source to an 80*68 machine
and again got the byte-orde of the client-ID wrong.  That's odd because
when I port code, I don't just unleash the result and assume it works.
I do things like test it in private with copies of the original on
the original platform.

Vernon Schryver

More information about the DCC mailing list

Contact by mail or use the form.