OK, here is how I architected it:

I had a shared hosting account with Hostway, accessable via POP3. I created 
10 email accounts on my hosting account. I set up James' fetchpop service to 
grab the email off the Hostway accounts every 5 minutes. I then set up my 
program to query the James server every 5 minutes as well. My program would 
collect all of the email from James, sort it, count the emails, do some 
simple word searches (e.g. how many emails contained the word "porn", 
"mortgage", "free", etc), and save every email in a text file inside a 
folder which was date stamped.

So it's basically: (Hostway servers collect email and put it into my 
account) -> (James server pulled down the emails via fetchpop) -> (My 
program grabs email off of James, sorts the email)

I was going to "sprinkle" the 10 email addresses around the web in highly 
selected spots, e.g. one email address would be frequently posted on a forum 
used by teenagers, another for adults, techies, etc. But the project just 
never got off the ground, even though I had acquired a domain name for it 
and the program was essentially done.



Personally, what I really wanted to do is this:

1. Grab a dedicated server or a better shared hosting account, something 
that allows a continuous Java thread to run... The shared accounts I could 
afford didn't allow me to run a continuous Thread, and demanded $20 extra to 
add a JVM. As a poor college student, that's out of the question. The server 
I had my eye on was about $79 monthly (free setup, plenty of bandwidth as 
well).
2. Set up James. Disable every service except for POP3. Lock it down nice 
and tight.
3. Set up a PostgreSQL install. Configure it so it uses plenty of space, and 
optimizes writes, not so much reads.
4. In the spoolmanager, delete every Matcher and Mailet pair. Add a single 
Matcher/Mailet. 
5. Use the Matcher "all", and immediately write all incoming messages into 
the database.
6. Every hour/day/week (depending on how many messages you get) a cron job 
starts up the email analyzer. This is the custom program I really wanted to 
write. Analyze every damn thing about the emails stored in the database. 
Figure out how many have SPF, spellcheck the emails, grammarcheck, the 
average length of the email, the average length of the word, readability 
stats, etc. 

I know this sounds hard, but it is really easy. It's just text, and every 
email message is a highly structured textfile, making analyzing easy.
7. Put the results into a webpage, put it for the world to see, and clear 
the db for the next set of incoming mail. 

I actually shopped this idea around my college, trying to get some funding 
for a dedicated server from some CS/Engineering slush fund or whatever, but 
there wasn't much enthusiasm for the idea. The Student Government wasn't 
much help either. Oh well.

IMPORTANT: IF ANYONE finds the above idea interesting, drop me a line. I'd 
gladly restart my project if someone/some company would foot the bill for 
hosting, or maybe even loaned me a server and bandwidth from their premises. 
It would make a great resume builder for me too.

Most of you can tell, I've put plenty of thought into this project already 
;-)

As for the code: I should have everything collected, including any help and 
ref files I used, by tomorrow. Give me a ring tomorrow evening if I haven't 
sent it out.

Thanks, (and sorry about the length of the email. I'm quite prolific when 
I'm discussing a project I'm enthusiastic about)

On 9/13/05, nurularesya ;p <[EMAIL PROTECTED]> wrote:
> 
> thanks..hopefully you will find the source...and thanks again...
>  then do spam honeypot will work in james?any suggestion? ;p
>  thanks,
> nurularesya
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around 
> http://mail.yahoo.com 
>

Reply via email to