[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] DELAY AND FORCE

20 minutes ago, John Cowan wrote:
> Andre van Tonder scripsit:
> > p 14: Waddell showed a long time ago that DELAY and FORCE are not
> >        the correct primitives for lazy evaluation in a strict
> >        language.  In short, it is impossible to write properly
> >        iterative lazy programs that don't have memory leaks with
> >        just DELAY and FORCE.  This is discussed in more detail in
> >        SRFI 45 and references therein.
> I agree, and proposed LAZY and EAGER as ticket #55.  However, the WG
> voted it down on the strength of Alex's arguments at
> https://groups.google.com/forum/#!msg/scheme-reports-wg1/qzAeT7395Sk/fTlT-BYTVysJ
> .  If you can rebut these, please do.

That thread is questioning experience with `lazy' and concludes that
it shouldn't be included in the base language.  One thing that has
been made very clear (by multiple people, and in many texts) is that
the naive implementation of `delay' and `force' is a bug when
considered in terms of its safe-for-space-ness.  Andre suggests
removing them in their current form from the language which IMO is at
least as justified as not including `lazy'.

          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

Scheme-reports mailing list