Update 9/4/2003: Please note that these somewhat cumbersome instructions below are now deprecated. You can now install SpamBayes with the simple MS Outlook plugin, provided by the SpamBayes project. See the download page for the plugin. The instructions below should only be followed if you want to install the latest bleeding-edge release of SpamBayes.
Ever since I last wrote about SpamBayes, I couldn’t wait to try to install it on my PC. The problem was that the Outlook plugin that was created for SpamBayes wouldn’t play nice with Outlook 2002 on Windows 2000–the exact configuration I’m running. Go figure.
Since then, I’ve been dying to install this software. The spam problem at work has upgraded from mildly annoying to prolific–spam now consists of 75% of my daily email. The majority of it is mild: cable boxes, money scams, and diet pills. However, at least a handful of these emails are for products that I’d rather not mention in a public forum. For months, everytime I open my inbox I’ve been anxious that a coworker will turn the corner and see me looking at (use your imagination). I’ve tried addressing the problem by setting up keyword-based rules in Outlook. They’re at the point where they take care of most of the problem, but I still get a handful of spam that leaks through.
On June 21st, the SpamBayes team released a third alpha version of (1.0a3) the software and, while they haven’t yet released a nice install package, I was able to download the source code and run an installation program in a Python GUI–all of which was relatively painless. If you’re beleagured by spam and feeling adventurous, give this a whirl:
# Download the the latest SpamBayes Zip file (e.g. SpamBayes-1.0a4.zip)
# Unzip the file and you’ll get a folder called “spambayes”
# Move this folder into C:\Progam Files
# Download the latest Python build (2.3)
# Run installer for Python 2.3, which will create a folder in your start menu
# Download the required Python Win32 Extensions for Python 2.3
# Run the installer for the Python Win32 Extentions–this program will install a Python GUI in your start menu (Note: you must have Python installed first!)
# Open Start → Programs → Python 2.3 → Pythonwin–that will open the Python GUI
# Click File → Open… and click to C:\Program Files\spambayes\Outlook2000\addin.py (Note, the path for version 0.4a is slightly different: C:\Program Files\spambayes-1.0a4\Outlook2000\addin.py).
# Click File → Run…, don’t change anything in the dialog box, and click OK
# The script will run and should not return any messages
# Open (or restart) Outlook–the install was successful if you have a new “SpamBayes” toolbar
After all of that is completed (whew!) you’ll want to open the SpamBayes Manager via the toolbar, and click on the About… button. That will give you some help on setting up the plugin.
Now, rather than dreading spam, I can’t wait to get more so I can train SpamBayes to kill it–bring it on. 
I’m still not convinced client side filtering of spam is the way to go, in fact, i think it is down right silly. My reasoning:
Email is about convience. It is about being able to quickly get a message to someone, so that you can get a ‘quick~ish’ reply. If you need an ASAP reply, you use the phone or an Instant Message Program.
To that end, reading email from only one computer seems rather … counter productive. If I am at work I want to see the same mail that I need see at home. If i am at home, I want to see the same mail i need to see at work. If i am on the road (work/vacation/otherwise) I want to see the same email there that i need to see at work and at home. I may have a network connection that I can plug my laptop into, and I may not.
Taking that into account, If i setup my mail to filter on one machine, I need to setup my mail to filter on all my machines. Granted, I have more machines to deal with than most people do (3 locations at work, 1 at home, and my laptop) … so, I will make it more general:
lets say you have a Windows 2000 Computer and check your email with outlook express 2002. Lets also say you hav a redhat 9 and check your mail with evolution. If you filter your spam with this dual booting machine, you now have to setup spam filters in 2 different places.
Not to mention ‘normal’ filters which filter your non-spam mail to the correct folders and whatever else people do with it.
That said, i believe strongly that email should be delt with completely server side, in such a fashion that one only has to check one email account to read /all/ their email regardless of where it is coming from. This is easily achievable using programs such as fetchmail, and several other tools which will retrieve mail from web-mail hosts… and then using imap w/ ssl to connect and check your email. This can be easily used in conjunction with spam-assassin and procmail to handle all your spam filtering needs.
To deal with the aspect of not having one of your computers to check mail with, webmail is available.
Now, I understand that this isn’t an option for everyone, it all depends on how strict your work is about letting you get your work email to another account … it also depends on having a limited amount of shell access on your server to setup the filtering tools. But, I think, over the next couple of years, we will see less and less pop usage and more server side email solutions.
At least I hope so. Ever try to recover email from a extremely irked user after their harddisk as crashed? Not fun to explain to them that all that corispondance is lost forever.
Altp.
Dude,
I doubt Kate would even blink an eyelash if she caught you reading an advert for breast enlargement pills…….even if you are a guy. Spam stinks!
Your cubemate,
Chris
Hi Mike. You make some great points, and I largely agree with you that the spam problem should be dealt with at the server level (perhaps ideally a client-server application that will get its training from clients, but manage email at the server). However, as you mentioned, this isn’t always possible.
I use Outlook 2002 to connect to Microsoft Exchange behind the firewall at work. That’s the only way I’m able to access email–no remote solution unless I VPN in from home. Due to this rigidity, I really needed something to work at the client level until corporate gets to the point where they’re capable of dealing intelligibly with this problem for a network-management standpoint.
SpamBayes is not the ideal corporate network administration solution, but for filtering my email for right now, I expect it to work wonderfully.
Mike, one additional thing I forgot to say re:
“Taking that into account, If i setup my mail to filter on one machine, I need to setup my mail to filter on all my machines.”
That’s part of the beautiful thing about SpamBayes–the filter setup is this:
1. Open manager tool
2. Indicate a folder full of spam
3. Indicate a folder full of good email
4. Click Train Now…
That’s it. Because SpamBayes does statisitical analysis, you don’t have to set up individual filters. Though you’d have to do this once for each of your four machines, you’d only have to do it once, and it’s a 5 minute process.
By the way, Day 1 with SpamBayes blocked all incoming spam this morning (about 10 pieces).
–
3. Indicate a folder full of good email
–
My folder for good mail is broken into well over 100 folders for good mail. On average, I probably receive close to 300 emails a day, about 100 of shich are spam.
1 folder for good mail would be insanely hard to deal with
. Does Outlook’s filters allow you to filter further after the spamfilters hit the piece of mail?
By breaking up my email (98% of which is work related), I can, at a glance, see which messages I need to deal with quickly (supervisors, deans, Users-problem-reports, etc …), and which can be ignored until lunch or after I go home.
and a side note: the bugtraq mailing list goes to my personal account but is needed by both my personal and professional lives. So, not all personal email at work is a bad thing … its the employees job to know when they can deal with what email at a given time.
that last statement could easily open up a whole new can of worms though. :-/
Sorry, I know i am probably beating a dead horse, or just mouthing off and annoying everyone … but this subject is one of the few (many?
) that I feel need addressed in the work place. Email is a huge time waster for a companies employees, but it really doesn’t need to be if it was handled in a logical fashion (email eats up about 30 minutes a day for me while i am at work, but i watch other people [including people in my office] spend more than 1/2 the day dealing with it).
anyway, now that I have wasted more than enough time on Ken’s website while at work today, i’m gonna go back to debugging my code
Altp.