Re: Email checking and filtering tool on Windows: call for help

You quote a disparate list of modules that suggest you don't have a firm idea of what you want to do and how you plan to implement it. But you did say you want to learn....

The modules broadly do:

# access POP3 mail, enumerate, download, delete
mail-pop3client 2.15

# validate email addresses
email-address 1.2
mail-rfc822-address 0.3

# sendmail to an SMTP server
mail-sendeasy 1.2
mail-sender 0.8.10
mail-sendmail 0.79

# mish mash of above
mail-box 2.054
[download]

The first thing to do is decide WHERE your tool is going to run, and how users interact with it. This will typically be either client side or server side, although a combo of clientside and server side is possible if you want to make it really complex.

If you are talking client side you need mail-pop3client Mail::POP3Client as this will let you pull mail down from a pop3 mailbox and delete trash. That is about all you can do client side as you can't really modify the emails in the POP box in any way, short of pulling them, deleting the copy in the POP box and then re-emailing them back into the pop box. You are constrained by the POP3 Protocol and what you can/can't do with it. You have to pull the email to analyse it so this is fairly inefficient. You should at least be able to mark emails in the POP3 box as read so you only pull them once.

If you are talking server side then you would typically use procmail to pipe the emails through your widget before they are delivered into the POP box. This is typically how SpamAssassin works.

I have no idea why you think email address validation modules might be of much use. By definition the to address is valid otherwise the mail would not be there. The from address may be undeliverable (due to being faked) but is extremely likely to be valid (from the RFC822/email syntax point of view). Even a moron spammer should be able to successfully fake a valid address.....

Similarly I have little idea why you are looking at modules that send email unless you are planning on challenge/response. Mail::Sendmail or Net::SMTP are fine.

Mail::Box tries to be all things to all users. It has the largest list of dependencies I have ever come across. You would want it to install with either PPM or CPAN as installing all the dependencies by hand is no fun at all.

What you need depends on what you want to do, and where and how you plan to do it.

In terms of research Google for keywords like 'Spam' 'Bayes' 'PopFile' 'Paul Graham A plan for spam' 'Gary Robinson spam' 'crm114' 'spam bayes' 'spamassassin test corpus'. You may find the preclassified corpus of emails from the spamassassin project useful for testing if you are looking at a Bayesian type approach or useless for challenge/response.

Good luck.

cheers

tachyon

Comment on Re: Email checking and filtering tool on Windows: call for help Download Code


The stupid question is the question not asked
	PerlMonks