[EuroPython] Talk database

Michael Hudson mwh@python.net
Thu, 01 May 2003 12:52:02 +0100

"M.-A. Lemburg" <mal@lemburg.com> writes:

> Michael Hudson wrote:
>>>>>Could someone please explain why the talk database is corrupted
>>>>>and all the work we've done on it so far gone down the drain ?
>>>Thanks :-/
>  >>
>>>Also, why did you start to send out emails to the speakers
>>>without asking the track chairmen for their consent ??? The
>>>emails you sent out look just completely silly... (do I
>>>sound angry enough ?)
>> (Please be a little careful slinging "you"s around.  It wasn't me who
>> sent the emails, and as far as there is an "organising cabal", you're
>> part of it too).
> Sorry, didn't mean you personally. I meant those who initiated the
> emails before informing the ones who have the direct contact to the
> speakers... explaining a mistake by an auto-generated email isn't
> exactly what people would expect after you've just cleared out
> their submissions !

No, this didn't go too well.

> W/r to constructive suggestions: I've asked for daily backups of
> the site several times before, Tom promised to look into adding
> them. 

OK.  I didn't know that.  Perhaps we should try to eschew totally
private communication in relation to the con?  One of the problems
we've been having is that e.g. you didn't know what was going on
wrt. the database, I didn't know you had tried to suggest backups
before, etc, etc.

The tracker is supposed to help here, of course.  Use the tracker,

> You could also try to revive the xmlrpc GetTalks/GetRegistrations
> scripts we used last year (I used them last year to at least save
> those two important databases in a separate location as backup).

I think this is being looked at.

There's that "you" again :-) One of the problems we're having this
year is that there are very few people (one?) who really understand
the web site.  I certainly don't.  We *have* to do something about
this before next year.

> And here's another suggestion: when hacking on the web-site,
> please please use a staging system *before* going into production.

I strongly suspect this suggestion is now redundant.  I hope so.

> Tim an I have tried to recover at least the data from our
> track: turned out that the browser had saved a listing with
> the titles, so at least that information is not lost.

Good!  I haven't found anything like that yet, but I don't think I've
lost too many talks...


  SCSI is not magic. There are fundamental technical reasons why it
  is necessary to sacrifice a young goat to your SCSI chain now and
  then.                                                  -- John Woods