Allen:

That construct only answered a single object, though one could build an array in some other way and answer it. It took a fragment of code, just like in a DoIt, evaluated it, and saved the result.

Dave


At 9:32 -0700 10/3/01, Allen Wirfs-Brock wrote:
>Gemstone was generally viewed as a specialized dialect rather than one of the "mainstream" implementations. My recollection is that their representative did not advocate for this feature.
>
>On a related issue, IBM Smalltalk had a feature (I think the syntax was ##(<statement-list>)) that created a literal array that was initialized from a set of Smalltalk statements that were evaluated at compile time(!!!!).  That feature was soundly rejected because of the object level coupling it created between the development and deployment environments.
>
>>Allen

Dave
-- 
_______________________________
David N. Smith
IBM T J Watson Research Center
Hawthorne, NY
dnsmith at watson.ibm.com