Rob,

Not so long ago (25 Aug 04) you wrote to me:
 >>  MK> Hm.  I am cutting and pasting what I have from FTS-0009 and am
 >>  MK> assuming that most, if not all, tossers follow:
 >>
 >>  MK>           ^AMSGID: origaddr serialno
 >>
 >> ^AMSGID: 1:249/303 05002dca 533.fidohusk 05002DCA
 >>          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~
 >>          origaddr in "syncronet network" serialno
 >>

 >> There's no requirement that origaddr should be enclosed in quotes if it
 >> contains spaces. It may be assumed, but it is not said explicitly, so
 >> above valid unless it is specified somewhere else 8-)

 RS> Sorry, I realize this is an old thread.
 RS> The break-down would be more like:
 RS>  ^AMSGID: 1:249/303 05002dca 533.fidohusk 05002DCA
 RS>           ~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 RS>           origaddr           serialno
 RS> or:
 RS>  ^AMSGID: 1:249/303 05002dca 533.fidohusk 05002DCA
 RS>           ~~~~~~~~~ ~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~
 RS>           origaddr  serialno      

 RS> The reason for the extra message-specific text (particularly,
 RS> "533.fidohusk" in this example) is that it's been proven time and
 RS> again that a 32-bit "serialno" is not a sufficient amount of unique
 RS> data to denote a specific message from a specific system. And Fido
 RS> MTAs that store the entire MSG-ID (or a digest thereof) will get a
 RS> truely unique MSG-ID for each message generated by a Synchronet
 RS> system. The repeated, altered-case "serialno" is just to
provide more
 RS> unique data for a "digest" algorithm, if one is used. If Fido MTA
 RS> can't handle the additional message-specific data (e.g.
 RS> "533.fidohusk") then it won't be able to handle the
repeated serialno
 RS> either.

But in this case you introduce violations of FTS standard which explicitly
defines syntax of MSGID kludge. This is bad.

WBR, Alexander

--- Самоед 2004.1.20