powermail-discuss Digest #2893 - Wednesday, October 15, 2008
Microsoft Exchange
by "Richard Hart" <[EMAIL PROTECTED]>
Re: Microsoft Exchange
by "Richard Davis" <[EMAIL PROTECTED]>
SpamSieve Statistics
by "Jeremy Hughes" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Michael J. Hußmann" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Michael Tsai" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Jeremy Hughes" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Jeremy Hughes" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Michael Tsai" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Michael Tsai" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Jeremy Hughes" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Jeremy Hughes" <[EMAIL PROTECTED]>
Re: SpamSieve Statistics
by "Michael Tsai" <[EMAIL PROTECTED]>
----------------------------------------------------------------------
Subject: Microsoft Exchange
From: "Richard Hart" <[EMAIL PROTECTED]>
Date: Tue, 14 Oct 2008 21:05:14 -0700
I have searched the list archives, but I cannot find any way to set up
PowerMail to access email on a Microsoft Exchange server. I know people
have done it, and that PM cannot send messages using an Exchange
address. But can it receive messages?
RH
----------------------------------------------------------------------
Subject: Re: Microsoft Exchange
From: "Richard Davis" <[EMAIL PROTECTED]>
Date: Tue, 14 Oct 2008 23:39:51 -0500
You should be able to POP3 Smtp and or IMAP BUT those services must be
turned on at the exchange server. They are off by default because
nobody would ever use anything except outlook or web would they
Sent from my iPhone
On Oct 14, 2008, at 11:05 PM, "Richard Hart"
<[EMAIL PROTECTED]> wrote:
> I have searched the list archives, but I cannot find any way to set up
> PowerMail to access email on a Microsoft Exchange server. I know
> people
> have done it, and that PM cannot send messages using an Exchange
> address. But can it receive messages?
>
> RH
>
>
----------------------------------------------------------------------
Subject: SpamSieve Statistics
From: "Jeremy Hughes" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 12:51:53 +0100
Is the Statistics window in SpamSieve 2.7.1 broken?
Mine shows:
Filtered Mail
0 Good Messages
0 Spam Messages
0 Spam Messages Per Day
SpamSieve Accuracy
0 False Positives
0 False Negatives
0.0% Correct
Corpus
2 Good Messages
11 Spam Messages (85%)
1,009 Total Words
Rules
6,354 Blocklist Rules
3,072 Whitelist Rules
Showing Statistics Since
14/10/2008 12:00
Jeremy
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Michael J. Hußmann" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 13:59:55 +0200
Jeremy Hughes ([EMAIL PROTECTED]) wrote:
> Showing Statistics Since
> 14/10/2008 12:00
Maybe you should set the date to a year ago or so. You are looking at
the statistics for just one day.
- Michael
Michael J. Hußmann
E-mail: [EMAIL PROTECTED]
WWW (personal): http://michael-hussmann.de
WWW (professional): http://digicam-experts.de
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Michael Tsai" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 09:34:11 -0400
On Oct 15, 2008, at 7:51 AM, Jeremy Hughes wrote:
> Is the Statistics window in SpamSieve 2.7.1 broken?
No. Only you can know whether the statistics shown match reality, but
most likely they do. What it says is that SpamSieve has not been asked
to filter any messages since yesterday at noon. This could be because
you didn't receive any messages or because of a setup problem in
PowerMail. Less likely, it's possible that a damaged database file is
causing SpamSieve to show fewer messages than there actually were--you
could tell if this were the case because the statistics would be
inconsistent with the log.
Secondly, it says that you've only trained SpamSieve with 13 messages
(total, not just since yesterday). Or maybe you recently reset the
corpus?
--
Michael Tsai <http://c-command.com>
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Jeremy Hughes" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 13:46:58 +0100
Hi Michael,
>> Showing Statistics Since
>> 14/10/2008 12:00
>
>Maybe you should set the date to a year ago or so. You are looking at
>the statistics for just one day.
It makes no difference.
I reset the corpus and deleted some rules earlier today.
But as you can see from my last email, SpamSieve is filtering mail (2
Good Messages, 11 Spam Messages) but failing to report this in the
Filtered Mail and SpamSieve Accuracy sections. This was happening before
I reset the corpus.
Are you using 2.7.1?
Does the Statistics window work for you?
Jeremy
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Jeremy Hughes" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 16:08:34 +0100
Michael Tsai (15/10/08, 14:34) said:
>> Is the Statistics window in SpamSieve 2.7.1 broken?
>
>No. Only you can know whether the statistics shown match reality, but
>most likely they do. What it says is that SpamSieve has not been asked
>to filter any messages since yesterday at noon.
SpamSieve is filtering messages fine - but it's failing to report this
correctly in the Statistics window (see below).
>This could be because
>you didn't receive any messages or because of a setup problem in
>PowerMail. Less likely, it's possible that a damaged database file is
>causing SpamSieve to show fewer messages than there actually were--you
>could tell if this were the case because the statistics would be
>inconsistent with the log.
>
>Secondly, it says that you've only trained SpamSieve with 13 messages
>(total, not just since yesterday). Or maybe you recently reset the
>corpus?
I reset the corpus earlier today, because it had got large and I thought
this might be causing the Statistics problem. The "Corpus" section
reports the number of messages that have been filtered since the corpus
was reset. In my previous email this showed as "2 Good Messages, 11 Spam
Messages (85%)". It's now up to "37 Good Messages, 74 Spam Messages
(66%)", but Filtered Mail still says "0 Good Messages, 0 Spam Messages etc."
I wonder if the problem is caused by the fact that SpamSieve's
History.db file has got quite large (122.9 MB).
Jeremy
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Michael Tsai" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 11:30:03 -0400
On Oct 15, 2008, at 11:08 AM, Jeremy Hughes wrote:
> I reset the corpus earlier today, because it had got large and I
> thought
> this might be causing the Statistics problem.
There is no relation between corpus size and the statistics.
> I wonder if the problem is caused by the fact that SpamSieve's
> History.db file has got quite large (122.9 MB).
The size doesn't matter, but it sounds like the file is damaged. You
could either start a new one (hold Command-Option when launching
SpamSieve) or send me the file (e.g. on an iDisk or Dropbox) and I'll
repair it.
--
Michael Tsai <http://c-command.com>
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Michael Tsai" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 11:04:44 -0400
On Oct 15, 2008, at 8:46 AM, Jeremy Hughes wrote:
> But as you can see from my last email, SpamSieve is filtering mail (2
> Good Messages, 11 Spam Messages) but failing to report this in the
> Filtered Mail and SpamSieve Accuracy sections.
The 2 and 11 are the numbers of messages in the corpus (i.e. the ones
you've trained it with). They don't show that it filtered any
messages. You'd need to look at the log to determine whether it did.
--
Michael Tsai <http://c-command.com>
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Jeremy Hughes" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 17:02:09 +0100
Michael Tsai (15/10/08, 16:30) said:
>> I wonder if the problem is caused by the fact that SpamSieve's
>> History.db file has got quite large (122.9 MB).
>
>The size doesn't matter, but it sounds like the file is damaged. You
>could either start a new one (hold Command-Option when launching
>SpamSieve) or send me the file (e.g. on an iDisk or Dropbox) and I'll
>repair it.
Thanks!
Resetting the history has fixed the problem :)
Jeremy
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Jeremy Hughes" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 17:10:06 +0100
Michael Tsai (15/10/08, 16:04) said:
>> But as you can see from my last email, SpamSieve is filtering mail (2
>> Good Messages, 11 Spam Messages) but failing to report this in the
>> Filtered Mail and SpamSieve Accuracy sections.
>
>The 2 and 11 are the numbers of messages in the corpus (i.e. the ones
>you've trained it with).
I didn't train it with any messages - these are messages that SpamSieve
has filtered (the current totals are 50 good and 112 spam).
It doesn't seem to be necessary to train SpamSieve in this case - it is
using its existing whitelist and blacklist and filtering quite
successfully with these. I've corrected one false negative out of 162
messages so far.
>They don't show that it filtered any
>messages. You'd need to look at the log to determine whether it did.
Yes - the log shows that they have been filtered.
Jeremy
----------------------------------------------------------------------
Subject: Re: SpamSieve Statistics
From: "Michael Tsai" <[EMAIL PROTECTED]>
Date: Wed, 15 Oct 2008 12:27:41 -0400
On Oct 15, 2008, at 12:10 PM, Jeremy Hughes wrote:
>> The 2 and 11 are the numbers of messages in the corpus (i.e. the ones
>> you've trained it with).
>
> I didn't train it with any messages - these are messages that
> SpamSieve
> has filtered (the current totals are 50 good and 112 spam).
You did, in a fashion, if you had auto-training turned on:
<http://c-command.com/spamsieve/manual-ah/auto-train-with-incomin>
When the corpus is very small, SpamSieve will auto-train with
virtually all the filtered messages. So the corpus will built up
automatically and you only have to manually train it if there's a
mistake.
--
Michael Tsai <http://c-command.com>
----------------------------------------------------------------------
End of powermail-discuss Digest
