ietf-nntp Backfill

[Chris Lewis]

In message "Re: ietf-nntp Backfill", 'sob at academ.com' writes:

>I agree with Nat. I got the impression that most folks were basically surprised
>by the issue of backfill. Erik himself said that it was always understood that
>article numbers were monotonically increasing.

I've always understood it that way too.

>Unless there is a strong desire to support backfilling explicitly, I think
>the new document should state that articles numbers are monitonically
>increasing and that backfilling is not supported.  It certainly makes
>server and client implementation simpler (I think).

If you allow backfilling, you essentially have to abandon article
ranges in .newsrc, and you might as well suggest that newsreader authors
switch to storing message-ids in .newsrc instead of article numbers.
While that would transparently solve multi-server sourcing of articles,
the resultant explosion in .newsrc size would be way beyond being
practical.

I consider backfilling (except for epoch reset) to be broken behaviour,
and it should not be enshrined in the standard.
--
Chris Lewis, Senior Network Security Analyst, Nortel.
clewis at nortel.ca; Dept 4C16, Ottawa, Canada.  (613) 763-2935.