William Albert wrote:
> Dr Robert Young wrote:
>
>> We have been using Razor2 for some time on SA 3.0.4. I was recently
>> reading about DCC. We have never tried it, so I was wondering about
>> opinions as to its use. How effective is it? Should it be used with,
>> or in place of, Razor?
>
> SpamAssassin will use both, so there's no need to choose
> between the two unless network traffic is a major concern.
I would use both. Razor2 is more effective than DCC, but DCC seems to
generate more errors in connecting, so perhaps it could have the same
recognition rate as Razor2.
There is no guarentee that either one won't produce false positives or that
it has an 100% uptime. So using both (and perhaps even Pyzor) for me seems a
sensible thing.
Some stats for recognized spam ( >10 ) over the last 10 days:
BAYES_99 ( 97%)
RAZOR2_CHECK ( 86%) <--
RAZOR2_CF_RANGE_51_100 ( 86%) <--
HTML_MESSAGE ( 69%)
DIGEST_MULTIPLE ( 66%)
URIBL_BLACK ( 66%)
PYZOR_CHECK ( 57%) <--
URIBL_JP_SURBL ( 57%)
DCC_CHECK ( 57%) <--
URIBL_SBL ( 54%)
URIBL_SC2_SURBL ( 51%)
URIBL_OB_SURBL ( 51%)
URIBL_XS_SURBL ( 47%)
Kind Regards,
Sander Holthaus
