http://bugzilla.spamassassin.org/show_bug.cgi?id=2894
------- Additional Comments From gary@intrepid.com 2004-01-03 10:50 -------
Since many SA users do not always communicate in English (including some
English speakers <g>), it would seem difficult to impossible to use a spell
checker without knowing the language of the sender.
Here's a different idea, ralating to the parsing of multipart/alternative
messages: process *only* the HTML part. Since we know that the spammer will
likely use the text part to discuss the spam in the HTML part, then SA should
focus on where the spam is likely to be: the HTML part.
In addition if sufficient work is put into ferreting out only the visible part
of the HTML (ie, ignoring text that is "invisible" to the reader, because it
is hidden in a font that is colored the same as the background, or there is
Bayes poison sprinkled in HTML comments bogus tags and such, and only the
visible part is passed to Bayes, and/or scanned by SA, then perhaps the
ability to spoof with a text part that doesn't match the HTML part goes
away, and so does a major source of Bayes poison.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
------- Additional Comments From gary@intrepid.com 2004-01-03 10:50 -------
Since many SA users do not always communicate in English (including some
English speakers <g>), it would seem difficult to impossible to use a spell
checker without knowing the language of the sender.
Here's a different idea, ralating to the parsing of multipart/alternative
messages: process *only* the HTML part. Since we know that the spammer will
likely use the text part to discuss the spam in the HTML part, then SA should
focus on where the spam is likely to be: the HTML part.
In addition if sufficient work is put into ferreting out only the visible part
of the HTML (ie, ignoring text that is "invisible" to the reader, because it
is hidden in a font that is colored the same as the background, or there is
Bayes poison sprinkled in HTML comments bogus tags and such, and only the
visible part is passed to Bayes, and/or scanned by SA, then perhaps the
ability to spoof with a text part that doesn't match the HTML part goes
away, and so does a major source of Bayes poison.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.