SpamPal vs. K9

[O.J.]

Hello!

I see that SpamPal is the recommended anti-spam utility on the 2003
PricelessWare list, but K9 has replaced it on the 2004 list.

I tried doing a Google search to find the pros and cons between the two
programs, but the only thing I came across was that SpamPal is slower and
uses more system resources due to the computer language that it's written
in.

So my questions are:

Why did K9 replace SpamPal on the PricelessWare list?

What are the pros and cons between the two programs?

Which is the better program?

Is there another program that is better than either of these two?

Thanks,
O.J.

 

[Bob Adkins]

I see that SpamPal is the recommended anti-spam utility on the 2003
PricelessWare list, but K9 has replaced it on the 2004 list.

I tried doing a Google search to find the pros and cons between the two
programs, but the only thing I came across was that SpamPal is slower and
uses more system resources due to the computer language that it's written
in.

So my questions are:

Why did K9 replace SpamPal on the PricelessWare list?

What are the pros and cons between the two programs?

Which is the better program?

Is there another program that is better than either of these two?

Good questions.

I have never tried K9, but I have used SpamPal for about a year. It's
faultless from where I sit. Easy to configure, easy updates, easy to install
plugins, etc. etc. I can't imagine it being any easier or more effective,
but my imagination may be lacking.

Bob

 

[badgolferman]

What are the pros and cons between the two programs?

Which is the better program?

Is there another program that is better than either of these two?

I started with SpamPal but have since migrated to K9. As you said, SpamPal
took up more resources and I had to get blacklist updates from wherever.

K9 has a simple interface, is easily taught what messages to block or allow
through, has a server check so you don't have to download messages and can
view them while still on server, you can have black and white lists, and has
a neat statistics page where you can see how well it's done. I have been
around 99.2% accuracy for a while, but mainly because there are certain
vendor advertisements I have been reclassifying as Good mail rather than
Spam. Keep in mind there is a short learning curve for the program to learn
what YOU consider Spam and what YOU consider Good mail.

There may be better programs around, but the simplicity and effectiveness of
K9 have made it a winner for me. Of course personal preferences are always
different and you should try these programs for yourself. I dare say if you
start with K9 you will not change over to SpamPal.

 

[Don Birdsall]

Hello!

I see that SpamPal is the recommended anti-spam utility on the 2003
PricelessWare list, but K9 has replaced it on the 2004 list.

I tried doing a Google search to find the pros and cons between the two
programs, but the only thing I came across was that SpamPal is slower and
uses more system resources due to the computer language that it's written
in.

So my questions are:

Why did K9 replace SpamPal on the PricelessWare list?

What are the pros and cons between the two programs?

Which is the better program?

Is there another program that is better than either of these two?

Thanks,
O.J.

Both are good. I started off with K9 and switched to SpamPal. K9
generated occasional errors. Bayesian filtering does not work well
for me. The 'good words' are the same as the 'bad words' just
arranged differently.

I don't think either one is 'better'. It depends on you. Use K9 for a
couple of weeks. Then use SpamPal for a couple of weeks. If the
choice isn't clear, flip a coin.

Don

 

[Aaron]

I've used K9 for best part of a year until about 3 months ago, when I
discovered SpamPal and have now switched to it. Although K9 is
excellent, in my view there is no comparison in that SpamPal has such
a multiplicity of types of filter (including Bayesian) and is
configurable to so many different user requirements. For example, you
can get SpamPal to check the input from Hotmail when you use Hotmail
Popper etc.

Well, if you know how to chain proxies, you can use hotmail popper with
anything from K9 to POPfile.

Either of these programs will blow the socks off anything else around,
but in my book SpamPal is the champ, and the more professionally
crafted of the two. Having said that, K9 is simpler for beginners to
understand and set up, but is more limited in what it can do.

Well maybe. But even k9 these days allow you to consult a open relay
list, much like spampal proper as well as supporting black/whitelists
with regular expressions.

So there isn't as much a difference between the 2 now. The focus is
probably different though, K9 banks on bayesian filtering with open relay
blacklists as backup, Spampal is probably the other way around? The last
time I tried spampal the bayesian module was in beta.

Similarly stuff like Frontgate mmx, provide the 3 main filtering
techniques

1) Bayesian
2) Open relay black lists
3) Normal bad words list (typically a regexp list)

Still if you like mutiple techniques/filters something like SPAMIHILATOR
is hard to beat.






Aaron (my email is not munged!)

 

[Aaron]

I used spampal for over a year, but got fed up with it downloading
entire messages, even if it could classify 99.9% as spam and they
would therefore be put in a spam folder or easily deleted.

Now on Frontgate MX, which while I have to approve it's
classification, it downloads only headers- so deletes stuff from the
server where it is clased as spam.

How exactly is it deciding what is spam just by looking at the headers?
It's just using RBL lookups? So it doesn't use the other filters?

If you can find a way for spampal to do this, i wouold go back to
spampal - as it is very automatic (once you trust it)

Hmm, there's no reason really why spampal shouldnt do this. Only if you
add bayesian filtering of the body, then you need to download everything.

Hope this makes sense

Aaron (my email is not munged!)

 

[Aaron]

SpamPal's Bayesian filter is tried-and-true. Farmer has updated it at
least 4 times, and it's very solid.

That might be so. But it's still an optional module. The focus of both
products is different.You want bayesian filtering mainly, than go for a
product that implements it as a core feature.

And don't believe what you read about the miracles of bayesian filtering,
it's nice, but it has disadvantages too compared to DNSBLs,

Bob

Aaron (my email is not munged!)

 

[Robin T Cox]

SpamPal vs. K9

I've used K9 for best part of a year until about 3 months ago, when I
discovered SpamPal and have now switched to it. Although K9 is excellent,
in my view there is no comparison in that SpamPal has such a multiplicity
of types of filter (including Bayesian) and is configurable to so many
different user requirements. For example, you can get SpamPal to check the
input from Hotmail when you use Hotmail Popper etc.

Either of these programs will blow the socks off anything else around, but
in my book SpamPal is the champ, and the more professionally crafted of the
two. Having said that, K9 is simpler for beginners to understand and set
up, but is more limited in what it can do.

 

[Bob Adkins]

So there isn't as much a difference between the 2 now. The focus is
probably different though, K9 banks on bayesian filtering with open relay
blacklists as backup, Spampal is probably the other way around? The last
time I tried spampal the bayesian module was in beta.

SpamPal's Bayesian filter is tried-and-true. Farmer has updated it at least
4 times, and it's very solid.

Bob

 

[Alastair Smeaton]

SpamPal's Bayesian filter is tried-and-true. Farmer has updated it at least
4 times, and it's very solid.

Bob

I used spampal for over a year, but got fed up with it downloading
entire messages, even if it could classify 99.9% as spam and they
would therefore be put in a spam folder or easily deleted.

Now on Frontgate MX, which while I have to approve it's
classification, it downloads only headers- so deletes stuff from the
server where it is clased as spam.

If you can find a way for spampal to do this, i wouold go back to
spampal - as it is very automatic (once you trust it)

Hope this makes sense

 

[Bob Adkins]

If you can find a way for spampal to do this, i wouold go back to
spampal - as it is very automatic (once you trust it)

I don't know if I would like to delete messages at the pop server, but it
would be nice if SpamPal did it both ways. The filters are very good. I'm on
2gb cable, so I don't notice the downloading. I bet I would be using
something else if I were on a slower connection.

I keep a log of all marked e-mails in case someone important can't get
through the filters. It rarely happens, but it has happened.

Can FrontGate MX keep logs of deleted messages?

Hope this makes sense

Oh hell yea. Makes perfect sense.

Bob

 

[Bob Adkins]

And don't believe what you read about the miracles of bayesian filtering,
it's nice, but it has disadvantages too compared to DNSBLs,

SpamPal uses both actually. I haven't checked to verify it, but I think my
DNSBL's cause SpamPal to flag far more spam than the Bayesian filter.

Ideally, the DNSBL's catch all the known spam headers, and the Bayesian
filter catch the unknowns... the ones that haven't made the DNSBL's yet.

Bob

 

[John Fitzsimons]

On Tue, 30 Dec 2003 03:33:17 GMT, Don Birdsall

Bayesian filtering does not work well
for me. The 'good words' are the same as the 'bad words' just
arranged differently.

< snip >

That depends on the program. That used to apply to K9. As PopFile only
classified on errors, and was therefore more accurate, K9 changed to
offer the same capability. I now find that both give me better than
99% accuracy in classifying spam.

I still think PopFile might be slightly better BUT if it is then we
are talking about a difference of less than 1% accuracy. I may also be
basing my opinion on past results. Now that K9 offers a DNS black hole
list as well they may now be the same.

Regards, John.

--
****************************************************
,-._|\ (A.C.F FAQ) http://clients.net2000.com.au/~johnf/faq.html
/ Oz \ John Fitzsimons - Melbourne, Australia.
\_,--.x/ http://www.aspects.org.au/index.htm
v http://clients.net2000.com.au/~johnf/

 

[Aaron]

SpamPal uses both actually.

Yes, as I acknowledged earlier. K9 does both too nowdays.

I haven't checked to verify it, but I
think my DNSBL's cause SpamPal to flag far more spam than the Bayesian
filter.

Well flagging more spam without any false positive is more important of
course.

Ideally, the DNSBL's catch all the known spam headers, and the
Bayesian filter catch the unknowns... the ones that haven't made the
DNSBL's yet.

Well I kind of like open relay black lists, mainly because you can just
download the headers without downloading the whole email and delete them.
Bayesian filtering is generally more accurate, but usually you need to
download the whole mail.






Aaron (my email is not munged!)

 

[Steve Basford]

Either of these programs will blow the socks off anything else around, but
in my book SpamPal is the champ, and the more professionally crafted of the
two. Having said that, K9 is simpler for beginners to understand and set
up, but is more limited in what it can do.

To help the situation a little, there is a auto-setup program being written to
help beginners setup SpamPal:
http://www.everydesign.com/products/spampal.asp

Having written the manual for SpamPal, I've tried to make things as simple for
the beginner as possible, whilst at the same time, provide power users with the
information they require.

I've tried K9 on another pc a week ago but found I really missed the extra
features of SpamPal and the HtmlModify plugin, which quite frankly... rocks!

Cheers,

Steve