Re: Test
  By: Michiel van der Vlist to Rob Swindell on Mon Jun 08 2020 10:38 am

 > Hello Rob,
 >
 > On Sunday June 07 2020 13:12, you wrote to me:
 >
 >  >> RS> Huh... that "53db7808" looks suspiciously like a Unix time_t value:
 >
 >  >> Yes, it looks like it. So?
 >
 > Things are not always what they look like.
 >
 >  RS> It's a known-to-be-broken method of MSG-ID serialno generation.
 >
 > I am aware that using the unix time stamp as the only way of generating
 > sequence numbers is not a good idea. See FRL-1029.

Huh. You seem to be very proud of that document. I see nothing in there about
using Unix time stamps for MSG-ID serialno's (what you're now apparently called
"sequence numbers"). Whatever.

 >  >> All my users are assigned a point number and post using that point
 >  >> number. No conflict.
 >
 >  RS> And none of those users could post more than one message within the
 >  RS> same wall-clock second?
 >
 > Irrelevant. I see to it that my users use up to date software tha generates
 > unique serial numbers.
 >
 >  RS> On my system, that kind of thing happens all the time (e.g. when a
 >  RS> users uploads an offline reader reply packet).
 >
 > A basic problem with off-line readers. I don't have that problem because I
 > do not use that system. The message id's from my users are generated on
 > their system by their point software. It does not matter in what time frame
 > they arrive here.
 >
 >  >> As for messages posted from the same address in the same second:
 >  >> see FRL-1029.002 4.8.
 >
 >  RS> Your system doesn't appear to be using the method described in
 >  RS> FRL-1029.002 4.8.
 >
 > Tings are not always what they appear.

Okay, cool then.

                                            digital man

Synchronet/BBS Terminology Definition #51:
NNTP = Network News Transfer Protocol
Norco, CA WX: 78.5øF, 9.0% humidity, 6 mph ESE wind, 0.00 inches rain/24hrs
--- SBBSecho 3.11-Linux