-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Sandy,
Thank you for your comments. I've given it some thought and I think I'm
making progress towards a good rule.
Sample text we want to avoid:
acxfic z fvddx n th r j eghqyxun hlchd rpk zgrxlkd
bqnkstszzwyoyh heljsi
We can try a series of several rules that test for different patters, all
found in this text, and give each a small score. Legitimate email might
be likely to hit some of these, but not very many. Whereas the spam above
is likely to hit many.
body CONSONANTS_7 /[bcdfghjklmnpqrstvwxyz]{7}/
describe CONSONANTS_7 Body contains 7 consecutive consonants.
score CONSONANTS_7 0.5
body NO_VOWELS /\s[^aeiou]+\s/
describe NO_VOWELS Body contains a bock of text without any vowels.
score NO_VOWELS 0.5
body ISOLATED_CONSTANT /\s[bcdfghjklmnpqrstvwxz]\s/
describe ISOLATED_CONSTANT Body contains an isolated constant.
score ISOLATED_CONSTANT 0.5
body ALL_CONSTANTS /s[bcdfghjklmnpqrstvwxz]+\s/
describe ALL_CONSTANTS Block of text with only consonants.
score ALL_CONSTANTS 0.5
I wonder if we can run a spell check on the text and use the number of
spelling errors as a measure for spamminess. Would that work? Again,
there is the problem of my legitimate PGP signature, and URLs and the
like. So we shouldn't set that score very high either.
On Fri, Jul 25, 2003 at 04:33:27PM -0500, Sandy S wrote:
> Sorry to say I tried something like this already - it hit every we got an
> email order with a credit card number in it, or any string of numbers for
> that matter. Even if you added numbers to the list (/[^aeiou0-9]/), it
> would still hit on your PGP signature below, on strings like "http:", etc.
>
> I'm playing with Chris's OBFU rules, but I'll keep watching the list for
> more ideas, too - I'd love to be able to catch these!
- --
Daniel Carrera | OpenPGP fingerprint:
Mathematics Dept. | 6643 8C8B 3522 66CB D16C D779 2FDD 7DAC 9AF7 7A88
UMD, College Park | http://www.math.umd.edu/~dcarrera/pgp.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (SunOS)
iD8DBQE/IbHsnxE8DWHf+OcRAtxqAJ4731cMk7McUkEOeA+Pnevgvu0oLACgj8Fu
oU+KTToamM1BXM+xvoAd/m0=
=8zJ2
-----END PGP SIGNATURE-----
-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
