On 08/26/2009 02:02 PM, Paul Cockings wrote:
> Dov Zamir wrote:
>> On 08/26/2009 01:23 PM, Paul Cockings wrote:
>>
>>> Sven Karlsson wrote:
>>>
>>>>> Aha. Yes. I was right. DSPAM 3.6.8. Have you considered updating your 
>>>>> DSPAM setup? 3.8.0 at least. DSPAM 3.6.8 does not offer you much to 
>>>>> improve your situation you currently are facing.
>>>>>
>>>>>
>>>> Can 3.8.0 be used in production? I was thinking of moving directly to
>>>> 3.9.0, but I'm unsure about the stability.... Users are already
>>>> calling and complaining about ham ending up in the spamboxes :)
>>>>
>>>>
>>>>
>>> Hi Sven,
>>>
>>> There's been about 200 commits to the source since Dspam 3.8.0 was
>>> released (over two years ago). The Dspam devel team are highly focused
>>> on releasing 3.9.0 within the next few months, but you can check out and
>>> build the latest code from GIT today. Many of us are running the latest
>>> code on our own production servers, with currently MySQL+Postfix being
>>> the most tested setup. You'll probably find more enthusiastic support
>>> for the the latest code over 3.6.8.
>>>
>>
>> But, Paul, do you really believe that a newer version of dspam will
>> solve his problems? I remember getting 99.x% accuracy even with 3.68. It
>> seems to me that the problem is with the way they're sporadically
>> retraining, not with a particular dspam version. IMHO, If they don't
>> change their methodology, their accuracy is not going to improve much.
>>
> I was addressing 3.9.x stability. If your thinking of moving from 3.6.8
> to 3.8.0, IMO you should forget 3.8.0 and either a) wait for new release
> (a few months) or b) upgrade now to snapshot. I can't see the point of
> spending time upgrading to 3.8.0 when we know we've made over 200
> improvements on that relaese - no one on this list is going to be
> enthusiastic about answering questions on 3.8.0 because a lot of that
> has been fixed - its old news. IMO 3.9.x is less buggy then 3.8.0.
>
> You are correct then Sven needs to re-think retaining. The latest code
> isn't going to cure the problem but might make things a bit better, esp.
> as some improvements to MySQL driver, much better handling of HTML email
> and all round general bug fixes.

Yes, of course this is true. Svwn, in any case dspam, although it is 
probably one of the best solutions available, it is not maintenance 
free, and may not be suitable in every scenario. The whole idea of 
shared groups in dspam came, I believe, from the need to centralize the 
maintenance when you cannot or do not want to trust your users to do it. 
If you cannot rely on the administrators to do it, then maybe dspam is 
not the correct tool for you.

OTOH, there may be a few possible resolutions to the problem (assuming, 
of course that you upgrade to the present version so we can continue to 
help you). Perhaps all suspected spam should be delivered instead of 
quarantined. With most modern servers, you can have the suspected spam 
delivered to a specific mailbox for each user, and then let the user go 
through it and move the errors into/out of this special folder. You can 
then use a cron job to do the corrections automatically for you.
Also, perhaps, if company policy permits, users can themselves retrain 
dspam. This is normally not a good idea since many users are clueless in 
this regard, however, if your administrators can't or won't do it, then 
maybe at least some users will get improved statistics.
I'm sure there are other possibilities that I haven't mentioned, perhaps 
someone else on the list can suggest a suitable methodology for you. 
But, if none of these solutions is feasible, perhaps you should be 
looking at a non-Bayesian anti-spam solution. Look at spamassassin, for 
example (never believed I'd suggest this on the list). It may do better, 
being a more static anti-spam solution than dspam.
>
> Kind regards
> Pc
>
> !DSPAM:500,4a95165b262671462969113!


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Dspam-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspam-user

Reply via email to