Re: short-circuit control forms (& 'long names are doom')

[Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)]

In article <Xns90C7A5EA1B78Dcharleshixsonearthli@207.217.77.21>, Charles Hixson <charleshixson@earthling.net> writes:

> Short-Cut Booleans:  Sometimes these are mandatory.  They rarely 
> do harm in current code.  But they may limit the optimization 
> that will be possible in the future.  I can quite imagine a 
> compiler that decides that a multi-processor should run the 
> logic tests in parallel, unless they were simple expressions or 
> short-cut forms.

There is no need to look to the future or restrict the venue only to
multiprocessors.  Some modern uniprocessor systems have chips that
are capable of "out of order" execution.  Compilers for such machines
must take explicit action to prevent this if mandated by shortcut
bookean expressions.