https://issues.apache.org/SpamAssassin/show_bug.cgi?id=2011

Kevin A. McGrail <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #4 from Kevin A. McGrail <[email protected]> 2011-05-06 21:11:14 
UTC ---
(In reply to comment #3)
> If this is ever implemented, then I suggest a counterrule to account for
> certain combinations which are actually valid, such as land-mobile and amateur
> radio callsigns.
> 
> I think that this covers what is needed:
> U.S. land mobile:  [kw][a-z]{1,3}\d{2,4}
> Amateur radio:     [a-z0-9]?[a-z][0-9][a-z]{1,3}  (international)
> 
> I do note that there may be certain combinations of "numbered words" which 
> also
> fit the callsign patterns.  Such could negate the effectiveness of the rule,
> but would be necessary to avoid falsing.

I think these extra reverse rules complicate things a bit since the rule talks
about counting per line more so than on a word by word basis.

So I think the tweaking of the rule to hit only when it gets to Spammy
territory is what is intended.

For example, in the example email, we have a 3 words with numbers/characters
and 15 without.  

Realistically, though, this type of rule is going to misfire on ALL kinds of
technical documents, phone numbers, addresses, international zip codes, company
names, etc.  Plus no one has actually written a rule or code to implement this
idea.  It might have merit but someone needs to spearhead it.

My immediate reaction is that this concept unlikely to be worth the time.

-- 
Configure bugmail: 
https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

Reply via email to