On 2021-11-17 at 03:59:38 UTC-0500 (Wed, 17 Nov 2021 09:59:38 +0100)
Philipp Ewald <philipp.ew...@digionline.de>
is rumored to have said:
UTF-8 emails SHOULD be base64 encoded.
Hmm most of our mails we get are not base64 coded... (with charset
UTF-8) but OK
Base64 encoding is only necessary if there are non-ASCII characters
used. UTF-8 is a superset of ASCII & it is normal for MUAs to not encode
more than needed. As a result, they often do no encoding or use
quoted-printable encoding for just a few non-ASCII characters instead of
base64, which encodes whole messages and has a 33% data size penalty.
So, to be more correct: MANY UTF-8 emails MUST be base64 encoded. No
US-ASCII emails MUST be encoded.
So any UTF-8 witch is not base64 should get a spam rating bacause IT
SHOULD be base64 coded?
never mind....
This reflects a widespread misunderstanding about how SpamAssassin works
and the rationale behind the rules.
SpamAssassin rules are not laws in any sense. They do not prescribe or
proscribe any action. They do not reflect any sort of moral or ethical
judgment. They do not express or define technical correctness. They were
created almost entirely by human guesses, across almost 20 years, judged
and validated by our RuleQA process which determines the publication and
scoring of individual rules within the whole set on a daily basis.
Whether a rule is published and what score it is assigned by default
depends solely on how that rule has *proven* itself useful in
discriminating between spam and ham. A rule does not exist because it
"should" but because it WORKS. We do not force-publish rules because we
think mail should or should not have particular attributes; we TEST, and
the testing determines what gets scored. No amount of rigorous logic
grounded in technical theory can override the judgment of RuleQA.
--
Bill Cole
b...@scconsult.com or billc...@apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Not Currently Available For Hire
