Saving money (fuel) with Spamcop.net reporting

[stan_qaz]

I've been fussing about saving money on my spamcop.net reporting for some time, there are a couple issues with the way that MW sends reports that aren't the most cost effective.

1. MW sends the message as retrieved (spam throttle lines or up to 999 lines if you scroll down) to spamcop.net for quick reports. Sending more than enough of the body (one or two lines) to keep the parser happy wastes fuel as for quick reports the body is discarded at the spamcop parser. This also results in sending more reports than needed due to the size limits on incoming spamcop messages to the parser. If the message was trimmed to just the header and 2 lines you could fit the number limit into the report rather than hitting the size limit first.

2. Mentioned in passing above, if you trigger the full message retrieve by scrolling down MW fetches up to 999 lines of the message, all fetched data is then reported which wastes fuel (aka MY MONEY!!!) not only on quick reports but on full reports where the body is used by the parser.

--------------
I've suggested adding a reporting throttle so you could set the number of lines sent by each reporting service you have set up to help with this. Others have suggested having the throttle be adjustable to more lines than the spam throttle setting causing MW to fetch up to 999 lines before sending the reports. That would not be useful for spamcop reports due to the spamcop parser limits but would be for other services.

With a reporting throttle available you could configure quick reports to send a couple lines, full reports to send 100 lines while still fetching 200 lines that appear to be optimal for Bayesian and FirstAlert detection.

----------------------
What I have done as a work around is added a second spamcop.net account (using a different login address) to my existing paid account but not purchased any fuel. This lets me send spam with a URL in the body to the full report address and take advantage of the lack of advertising and delay screens and send spam with no URLs in the body to the unfueled account as quick reports where the delay and advertising screens are not encountered.

You do have to go through the mailhosts configuration again, test it and request that quick reporting be enabled, a bit of hassle but the admins showed no hesitation in allowing quick reporting on an unfunded account.

----------------------
The above will help out my budget a bit (the reporting throttle would help more as I could go back to fetching 200 lines) but it has left me with a problem, my laziness is now in conflict with my cheapness. Now I have to sort through the spams, look at the bodies while being EXTREMELY** careful not to scroll too far and fetch unwanted lines looking to see if the spam has a URL and needs full reporting or doesn't and can be processed through the quick system.

So now I guess I am looking at having to go back and duplicate all of my spam detection filters so that I can have versions that grab spam with URLs and versions that grab spam without URLs. That and grouping by filter should put the spams into easily reported blocks.

The downside is that I'm going to have to figure out just how to do this efficiently, some of these filters have a lot of rules and most look at the header only. Adding a rule to search the body for "http://" to each is going to be a pain and slow things down.

My question coming out of all of this is: Is there a better way to do the filtering that I'm missing?

=======================
** The only fix for fetching too many lines is to clear the account cache, exit MW to clear the memory cache and start your marking process over from scratch, a huge time waster!