henry (03/29/83)
The problems John Levine describes do occur; I can testify to the arrival of extra X-files due to line failures during X-file transmission, and the rest sounds fully believable. The sequence number wraparound is also something I've observed in our link with decvax; in fact, it's been a headache for decvax neighbors for some time, whenever a site doesn't connect for a few days for some reason. That's why decvax uses hexadecimal sequence numbers, to try to reduce the problem. In the particular case in question, I think the bulk of the problem is due to sequence-number wraparound. The distinguishing feature is that the problems John describes tend to hit one file at a time, while sequence-number wrapping gives whole bunches of orphaned files, which is what Steve described. Henry Spencer U of Toronto