> ISO-8859 is Latin-1, which is the standard character set and > there is no need to be encoding Latin-1 except to get around > content filters.
You're right. Testing with Outlook 2003 and some messages containing legit special characters I can confirm that all legit messages are "Quoted printable" encoded with =?ISO-8859-1?Q?... The subject lines of spam messages usualy are BASE64-encoded: =?ISO-8859-1?B?... But I've found also several legit cases where the e-mail client has base64 encoded the entire subject line or also only the word that contains a special character. (Some of them was send from a hotmail account. During bussines time the ratio between ISO-8859-1/base64 encoded legit and spam messages on our server is around 35/65. Note: We process a lot of messages in German and Italian. Also messages written in French or Spanish can contain special characters like ����������� So maybe it's a good idea to give some points for "=?ISO-8859-1?B?" but not too much to avoid FPs. > ... though I believe that > SpamChk will do decoding... Right. It will decode both quoted printable and base64 encoded subject lines before checking for keywords. At the moment it will not write the decoded string in the spamchk logfile but I think this will be changed in the next release. Markus --- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)] --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
