On 2018 Feb 06 00:13:38, wiki.synchro.net wrote to All:
Wiki edit of http://wiki.synchro.net/faq:misc by digital man:
[FTN MSGID] more rant
FWIW:
5. It's obviously still a problem if 2 different systems generate the
same message-ID serialno value, unless there's some other source of "uniqueness" in the Message-ID...
that other source of uniqueness is why the address is used in addition to the serial number... this is why separating the fields is a BadThing<tm>... the MSGID should be read and treated as a string... this is why stripping of trailing spaces has caused problems...
eg:
MSGID: 1:234/567 12345678
MSGID: 1:432/765 12345678
those are unique and distinct MSGIDs /because/ of the addresses...
MSGID: 1:234/567 12345678
MSGID: 1:234/567.0 12345678
these, too, are unique and distinct MSGIDs... granted, we generally drop the ".0" but in this case, and since it is all a string, it is certainly a different MSGID than the one above it... in fact, the following are all separate and distinct MSGIDs even though they all depict the same FTN system address in various forms...
MSGID: 1:234/567 12345678
MSGID: 1:234/567.0 12345678
MSGID: 1:234/567@fidonet 12345678
MSGID: 1:234/567.0@fidonet 12345678
granted, the address and the eight character hex serial number combination will
loop after 4,294,967,295 posts (if all the numbers are used) but that does fit the "three year" strategy to a point... no one ever really expected, at that time, that there would be systems archiving messages and storing more than three years worth without deleting and purging... certainly no one ever really expected that BBSes would be gating and importing millions of usenet posts when
usenet was available as a separate entity and those posts could stand on their own in their own format in their own message bases in the BBS
i thoroughly enjoy the part where you saw and recognized the word "should" in the definition of the address portion... that is the key to this whole thing and you are correct... i (finally) saw it years ago when you and i initially bumped heads about this... one only need look at other FTSC documents to see how "should" is used... specifically FTS-1006 which defines what words like "should" and "must" mean... now, maybe some folk will BTFO and actually read the specs they follow instead of interpreting them into what they think they say...
http://ftsc.org/docs/fta-1006.002
ya done good!
FWIW2: a lot of these problems have been around for years and years and years... we (TINW) just never/rarely ever saw them because there was so so so much traffic to wade through... now that the traffic level has dropped, it is much easier to see the problems...
FWIW3: yes, numerous folk did use a simple random number generator... others also used a simple t_time... both do present their problems and, in fact, a simple monotonic counter is pretty much what one should use... this was alluded
to by some folks back in the heydays of fidonet when discussions talked about looking at how internet emailers created their Message-Ids...
:APPLAUSE: :APPLAUSE: :APPLAUSE:
)\/(ark
Always Mount a Scratch Monkey
Do you manage your own servers? If you are not running an IDS/IPS yer doin' it wrong...
... If you're unhappy, you can always change your mind.
---
* Origin: (1:3634/12.73)
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net