<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>HTML notepaper</TITLE>
<META http-equiv=Content-Type content="text/html; charset=us-ascii">
<META content="MSHTML 6.00.2800.1400" name=GENERATOR>
<STYLE>BODY {
        MARGIN-TOP: 18pt; FONT-SIZE: 10pt; MARGIN-BOTTOM: 20pt; COLOR: black; FONT-FAMILY: Century Schoolbook; BACKGROUND-COLOR: aliceblue; TEXT-ALIGN: justify
}
A {
        COLOR: black; TEXT-DECORATION: none
}
A:active {
        COLOR: #660000
}
A:hover {
        COLOR: #660000; TEXT-DECORATION: underline
}
P {
        MARGIN-TOP: 8pt; MARGIN-BOTTOM: 8pt
}
Hdr-Ftr {
        COLOR: #004040
}
P.Hdr-Ftr {
        FONT-WEIGHT: 700; FONT-SIZE: 10.5pt; FONT-STYLE: italic; FONT-FAMILY: Bell MT,Calisto MT,Goudy Old Style,Century Schoolbook
}
</STYLE>
</HEAD>
<BODY>
<DIV dir=ltr align=left><SPAN class=703254117-06022004>SpamBayes only trains on
received messages, and it has a check to prevent you from training on a message
that you sent yourself. This is the check that caused the "No filterable items"
popup, but the warning would certainly be more helpful if it could be a little
more detailed about why the message was not "filterable". I'll look into whether
or not we can do that, and we should be able to at least include some
<EM>possible</EM> reasons why it would occur even if we don't know the exact
one.</SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=703254117-06022004></SPAN> </DIV>
<DIV dir=ltr align=left><SPAN class=703254117-06022004>As far as the need to
"Recover" it, I doubt that SpamBayes trained on it at all, so there shouldn't be
any mis-training issues caused by the fact that you can't train it
manually.</SPAN></DIV>
<DIV> </DIV>
<DIV align=left><FONT face=Arial>-- </FONT></DIV>
<DIV align=left><FONT face="Courier New">Kenny Pitt</FONT></DIV>
<DIV> </DIV><BR>
<BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px">
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma><B>From:</B> spambayes-bounces@python.org
[mailto:spambayes-bounces@python.org] <B>On Behalf Of </B>Coe,
Bob<BR><B>Sent:</B> Friday, February 06, 2004 8:23 AM<BR><B>To:</B>
spambayes@Python.org<BR><B>Subject:</B> [Spambayes] Funny behavior of Outlook
plugin<BR></FONT><BR></DIV>
<DIV></DIV>
<P class=Hdr-Ftr
style="FONT-SIZE: 14pt; MARGIN-BOTTOM: 0pt; FONT-STYLE: normal; FONT-FAMILY: Georgia,Imperial BT">R. K. Coe</P>
<HR style="MARGIN-BOTTOM: 2pt; COLOR: #660000" SIZE=1>
<P><SPAN class=444280013-06022004>I don't know whether this is a bug or a
feature; but it was a bit unexpected, so I thought I'd pass it
along.</SPAN></P>
<P><SPAN class=444280013-06022004>I use the Outlook plugin (0.85) with Office
2000 under Windows XP Pro. Last night I received a real message that got
misclassified as spam. I unthinkingly answered it before moving it back to my
inbox, so my answer went into the spam folder. (Outlook does that if you
answer a message in a folder other than your inbox.) So I selected it and
clicked the "Recover" button, whereupon I got the "No filterable mail items
are selected" popup, and the message didn't move. I was able to move
it without using Spambayes, and the Recover button did, of course, work
on the original message.</SPAN></P>
<P><SPAN class=444280013-06022004>My guess is that this behavior is a side
effect of an implicit decision not to assign any training value to a message
that hasn't been scanned. That's not illogical, but I don't think that
particular error message (whose ubiquity belies its lack of information
content) quite conveys the reasoning.</SPAN></P></BLOCKQUOTE></BODY></HTML>