Wishlist 2, going personal

You have suggestions to improve PopMan? Here is the right place to mention them.

Moderator: Christian

Post Reply
Posts: 7
Joined: 17 Dec 2007, 18:05

Wishlist 2, going personal

Post by Bas »

- You can use M. Serhat Çinar's wab2txt 0.01 (perl script/exe), to convert a OE wab file in a whitelist. (I've succesfully tested it.)

* a score for a spam probability; for each badword a number (I've read this somewhere else)
Some scores/cases (not all) may be save to delete, or put in a 'sure to delete' bin.
Initially a badword has a standard zero value to begin with. Not all badwords need to be weighted; colour addressing needn't be necessary, and perhaps a spam_regex notification can be replaced by the actual found word. I have defined a regex:string with several |-signes. Please see http://www.regular-expressions.info or http://zez.org/article/articleprint/11 for more information. Other regex related tools: 'V-Grep', 'Agent Ransack', 'RegexBuilder' (2 3) or perhaps 'Regex Assistant' (regex_tips.htm; could be useful for that 'txt markup' idea..)

* a timer/countdown may be of use; if you're about to hit the 'empty trash' button, it could happen, with a fast internet connection, the programme just did another mailcheck on the server, possibly deleting legitimate emails. Ghosting the 'empty trash' button around that time may not be required/mandatory. viewtopic t=238

And a few more, (also;) for discussion,
* when email is suspicious/verdächtig:
Message contains the following unknown words (dictionary check (Aspell), not found in dictionary) and
Message contains the following links (web and mail addresses) viewtopic t=358

* And, if suspicious; Before spammers turn to "insinuations" and "stories", messages with a weird layout could perhaps be relfected by an equally weird character show, as summed up (indicated/displayed) in Popman's rules column. This requires a simple (as in non-complex, and very much lighter than Flash/Silverlight) opensource (txt?-) animation package. /As instructed by the structure of the received message. Or something like this. Fun! If kept simple and light perhaps. Perhaps a hint then; do we have spam backfiring?! Replace "Mouse" with "badRat" ;)

* Message originates from outside; [domain], [domain country] (as found in the header, x-from/return-path, could be a victim mail address, but it identifies a suspectible message anyway)
viewtopic t=380 (Tommy's request; is it a good idea to include ".. and contains none of Good_Domains"? Perhaps it's good to count the rejected domains in a log as well, and keep them organised.)

* Message originates form a trusted domain, namely [domain] (from domain whitelist?) (Suggested being displayed in short: "orig: [domain]") The argument being that good or reasonally good domains won't let themselves being abused easily. And, some domains can't be abused at all because of their internal (Apache? webserver) usage restrictions?
This could be mentioned in the unfiltered message list Rules column (which as yet remains empty?)

* I also think the Rules column should show legitimate (known, whitelisted) senders, or mention that the sender is whitelisted

+ Ahem, with the risk of indeed overloading the internet (as it is already):
autoreply with: please include one of the following keyword in your message: [keyword1, keyword2]
- victims (should) get notified they're being abused (so take action, ask your internet provider for help)
- spammers may address/target you more specifically using this information, but I think they're also more easily dealt with that way. Anyone for comments on that?
(messages containing those keywords can be processed separately)
((suspect messages could be auto-replied with fake keywords)) -> Please discuss, or dismiss

+ A web repository storing legitimately email addresses, voluntairy, and to get a 1 to 5 score with 1 being victim 2 being unknown and 5 being strong, trustworthy/respected (and I would join/register with my email address, probably eventually getting a 3 or so)
(and why not, when you're already victim of spamming? Those email addresses are ours period.) (And yes, I know spammers use victims email addresses; please discuss, or dismiss. Please use arguments, the computer does too;) A 'No unsolicited advertising/werbung'-tag may not be required.

* I don't know if I'm right, but I like the fact your programme opts not to make use of a leaning mode, as it may (the reason being) too processor/processing demanding

Happy Spam analyzing, congrats, we could learn a lot from this (so thank you one hell of a lot, vielen Dank;) I know there are many priving eyes on the net keen on visiting this site in particular, and right they are.
    Posts: 7
    Joined: 17 Dec 2007, 18:05

    Online spamwords list database

    Post by Bas »

    consult an online internet database for BADWORDS (non-regex, but may be derived of which) entered/contributed by its users. The user gets a point for every other user who has entered the same BADWORD, and gets further points if he or she has the required number of the most common B@DWORDS on that list inquired just about right. Bear with me; that can either be 2^0, 2^1 and so on through 2^7 (so 1, 2, 4, all the way to 512 WORDS at most, I think, of which you would want to inquire from that list.)

    The user of PopMan for example can choose not to consult that list, choose to consult the top 32 BADWORDS or so (option 5), let the community decide the preferred amount for you, or go by the advisory of the respected/highest rated contributing person identified with it.

    Of course spammers might join too. So what are the consequences? We'll see, but it could be a step in the right direction, if it doesn't blow up altogether.. in which case the system stops working, and a more trustworthe may arrise.
    Posts: 7
    Joined: 17 Dec 2007, 18:05

    Third installment bit

    Post by Bas »

    If you'd allow me, it's as far as I'll go (it's relevant);

    Perhaps we also should analyse the first 'moonlanding'. (Everything was scripted?)

    It's always good to keep a weary eye on things (in a good sense that is). So that'll be it then.
    Yours sincerely,
    Bas Schoutsen
    Post Reply