Location: PHPKode > projects > ObsceneClean > SpecialRulesData-DATA-LAYOUT.txt
This file supplies the data to be used by the special rules. The data specifies 
what mitigating chars or words to look for and where. The data is rather uniform.

6 fields:

OW			Offensive Word
Rule type:	1= NEAR by chars (not words),'fag' following within 30 chars of 'smoke'
			6= NEAR any from a specified list of words or names 
			10 = if word preceded or receded by specific letter then reduce by weight x%, crap receded by 's' means its the gambling 'craps', but this would other mitigating factors
Near word 	Mitigating or aggravating word			
Proximity 	How close must it be? Use this as a measure: Average word length is 5 chars plus one space = 6 chars per word
Direction 	Search for near word. (2=BEFORE, 1=AFTER, 0=BOTH), 
Weight% 	Added to detected OW's probability. Weight is negative num if mitigating or positive if aggravating

so:  word, rule type, near word, proximity, direction, weight


SPECIAL RULES ARE:

21		AsianNamesGeneric - Is an East Asian Name?
1		Is the mitigating or aggravating word near OW by X num ver of chars?
22		Are CJK chars present?
25		Is Vietnamese language ?   Vietnamese uses part of roman alphabet.
23		Are only japanese chars present?
6		Does a first name precede the OW?
10		Is word preceded or receded by specific letter? Crap receded by 's' means its the gambling 'craps', but there could be other mitigating factors
20		Is an East asian first name? Is this rule plausible?
16		Are erotic words present?
Return current item: ObsceneClean