[plt-scheme] Re: [Larceny-users] side effects in R6RS modules

On May 5, 2009, at 2:47 PM, Robby Findler wrote:
> I do not think that "smart" is the right label to apply
> to a compiler that eliminates code whose effect is observable.

Correct if there are guarantees by the language regarding
whether or not (or how many times) such observable effects
are performed.  In the case in question, none of this is
specified: a system can invoke such library 0, 1, or more
times without being, strictly speaking, incorrect.

Aziz,,,