[SpamCop-List] Re: Can SpamCop be improved to handle non-English messages?

[Frank Ellermann]

Patto wrote:

> http://www.spamcop.net/sc?id=z796577385z87fa23e572278be28e1c21c5500232cbz
 
> Whatever they do, they will not be able to read anything if
> the original spam message is not in a Western character set.

Depends on the abuse desk, if it's in Korea I'd expect that
they can handle Korean character sets as well as the Korean
subset(s) of Unicode.

> Japanese (as this sample), Chinese, Russian texts are
> implicitly replaced with question marks.

That was something on your side, before you reported it, if I
understand the tracker correctly:

| Content-Transfer-Encoding: 7bit
[...]
| Content-Type: text/plain;

No charset in the Content-Type, the default for text/* types
is us-ascii, and that's also suited for 7bit.

> Wouldn't it be better if SpamCop would forward the original
> message intact, with the original character set encoding?

Depends on the charset, for UTF-8 and some others I'd say yes.

A real MIME-forward as message/rfc822 (Windows users like Mike
claim that this is an "attachment", because that's what their
popular mail agents do) should work in theory, but in practice
many abuse desks cannot handle minimal MIME conformance.

The next step would be multipart/digest, the first part copies
the header as text/plain, the second part is again the complete
spam (header+body).

But this also doesn't work with ignorant abuse desks for the
same reasons, some are not minimally MIME conformant, others
filter all multiparts, the works.

Maybe SpamCop could offer an opt-in service for abuse desks,
"yes, we can handle MIME and want it".

                         Bye, Frank