On Wed, 5 Mar 2025 15:18:43 +0100 Tom Hendrikx <t...@whyscream.net> wrote:
> Interesting to see all the variants and diacritics used. Maybe we can > improve some rules based on the variants. I never received anything > like this, so sharing for the people interested. I received some spams like this, a couple of years ago maybe? They were correctly identified as spam but I do not remember if the basic rules were very efficient. Maybe by the network tests or the Bayesian filtering and/or CRM114 -- these filters are also fed by my spamtrap addresses, this could be the reason for the detection. I agree that if the language is English, too many of these diacritics are a warning sign. Unless too English speakers are talking about some other language... Only trigger the filter if a significant percentage of letters have these diacritics signs? Or do not set a score too high?
