->> You better sit down. You are *voilating* the MSGID spec. And
->> probably *causing* dupes.

 JM> And you're 'probably' wrong.  Probably is just like 'assume'

no, not this time he is not... your misunderstanding of what a CRC is and
assuming that an 8 character HEX serial number is a CRC is where you have
gone wrong.

 JM> Leonard. Only this time, you can't count ME into the equation.
 JM> Something you might not be aware of is that WILDMAIL! and
 JM> ViaMAIL! combined make up some 10,000+ nodes (yes, that's
 JM> 10,000+ nodes) in FTS compatible networks.  Let's see, that's
 JM> potentially terabytes of mail daily, I'd say that makes for an
 JM> ideal proving grounds, don't you?

sure we can count all that and also add in that all these systems are
running FLAWED mail managers...

->> THERE IS NO "CRC" IN A MSGID!!

 JM> Okay, lets see if I can put 2 and 2 together to come up with the
 JM> correct answer for which the likes of you, seem to need.  Bear
 JM> with me, you might need to take it slow in order to digest it all.

are you ready for a LARGE plate of crow?

 JM> 8 digit serial number. In PASCAL we have what's called a LongInt.
 JM> It's a 4 byte data structure that when converted to a HEX string
 JM> just happens to be 8 characters in length. Hooray!, Okay we now
 JM> have the data TYPE.

this is only a coincidence... maybe you'd like to read something that shows
a viable alternative to CRCs?

 JM> 32 bit CRC (ie: LongInt) of the message contents, including origin
 JM> line, certain portions of the message header is a darn good 'serial'
 JM> number of the associated message.  I know of no better way to come
 JM> up with that information, do you?

CRCs are NOT unique... more than one string can give the same CRC value!
the serial number is NOT a "'serial' number of the associated
message" ... no way, no how... it is simply a number...

 JM> Ever wonder why file transfer protocols use a CRC to block check
 JM> data, because it produces a U-N-I-Q-U-E number. What a concept.

wrong... they use the CRC to determine of there was an error FOR THAT BLOCK only...

read my next message for one method of coming up with a serial number for MSGIDs...

)\/(ark