[seqfan] Seqfan Archive fixed (was : Is the Seqfan archive permanantly defunct?)

[Olivier Gerard]

Dear David, and all other members.

The archive system of the seqfan mailing list has long been more and more
defective
(around May 2018 as David writes).  There was a series of unfortunate bugs
between
the OS, python, mailman, pipermail, language settings, etc. who corrupted
important
files and skipped many messages.

I don't want to enter into the technical details but one of the challenge
of fixing it is that there many links on the web pointing to this archive
and that most
easy fixes would have broken all these links.

The system has been fixed now for a few weeks on a backup server and I have
decided
to use January 2020 and these recent requests as a launch point.

All mail archive prior to July 1st 2018 are from the ancien system and will
stay.

Unfortunately, all links copied in the OEIS  and other web pages pointing
to the archive
for mail posted between August 1st 2018 and today January 15th 2020 will
almost certainly be broken
and must be corrected because the numbering is different.

YOU MAY NEED TO RELOAD PAGES with your browser to see changes.

I am sorry for the inconvenience. The time needed to fix all the
consequences of the
bug was very long on an old and unsupported server.

I will happily help anyone about this topic.
Please post PRIVATELY to me, not to the list for any specific request.

Olivier GERARD
Seqfan Mailing List Administrator

On Wed, Jan 15, 2020 at 6:10 PM David Seal <david.j.seal at gwynmop.com> wrote:

> My question is prompted by Thomas Baruchel's comment in another thread:
>
> > I recently asked about the Binomial Modulo 2 transform and got a nice
> answer
> > from N.J.A. Sloane; I would have liked making some reference to an
> archived
> > version of his reply, but strangely I couldn't see it in the archive of
> the
> > mailing list; I thank him again. ...
>
> Comparing the numbers of Seqfan postings in my email files and in the
> Seqfan archive at http://list.seqfan.eu/pipermail/seqfan/, the latter
> matches the former closely enough (i.e. allowing for possible counting
> errors and time-zone-induced discrepancies) up to and including April 2018,
> but is under ~10% of the former from May 2018 onwards - for example, I've
> received 65 Seqfan emails so far this month, but this month's archive
> currently contains only 3 of them. I.e. it looks as if the archives stopped
> archiving the vast majority of Seqfan mails (90%+ of them) in May 2018.
>
> This was noted at the time (in a posting that didn't get archived...) and
> a reply (ditto) suggested that the problem was being worked on and the
> archive would be rebuilt when the work was done. Is this still the case, or
> has fixing the problem been abandoned?
>
> Whatever the situation is about this, I would suggest that an explanatory
> note after "To see the collection of prior postings to the list, visit the
> SeqFan Archives." in http://list.seqfan.eu/cgi-bin/mailman/listinfo/seqfan
> would be a good idea. Without it, people who encounter Seqfan for the first
> time and take a look at the archives to see what it's like are liable to
> get a very misleading impression of a pretty inactive list whose contents
> are highly disconnected...
>
> David
>
> --
> Seqfan Mailing list - http://list.seqfan.eu/
>