Speaking of "standard logic", your "answer" to the "question" actually contains a potential bug due to improper handling of NULLS; the COALASCE, as coded, may still return NULLS in the event that ALL fields contain NULLS and therefore needs an extra NULLs test: eg.

SELECT NULLIF(COALESCE( ..., ..., ...), [...]) AS ...

Or was it your intention to allow NULLs in the result in any event ? If not, let's skip any more talk about "standard developer gaussian distributions" (sic).

>A simple question:
>implement a N levels of defaults surrogate values:
>

>SELECT COALESCE(myField,Surrogate1.myField,Surrogate2.myField... ,SurrogateN.myField) AS myField
>FROM MyTable JOIN Surrogate1 ON .. JOIN Surrogate2 .... JOIN SurrogateN ON ...
>

>
>Now, implement this without NULL:
>- you must use a multiple CASE .... and the SELECT it becomes unreadable
>- you must remove a domain value for define a unknown value ( like emptys in VFP )
>and all the symmetry of the program it goes is made to fry
>
>Jim, not use NULL is the standard logic for a standard developer gaussian distribution.