From:
srenner@... [mailto:
srenner@...]
>If Joy is understood as linear (currently a topic of debate here),
It undeniably can be made so, if it isn't already.
>then is a Joy computation reversible? My intuition is that Joy can be
>made reversible if "garbage" is not collected during a computation
>but pushed onto a "garbage stack".
>[2 2 +] i == 4
>3 1 + == 4
>2 sq == 4
>123 121 - == 4
>Given 4 on the stack, what information do we need to reconstruct the
>computation that created it? Isn't this information exactly the same
>as the "garbage" we would normally collect?
I don't think so -- I think you'd need to have (for example) + return both
the answer and one of the addends. So
[2 2 +] i == 4 [2 2 +] ( ? )
3 1 + == 1 4
2 sq == 4
123 121 - == 121 4
I'm not sure about how to handle the quotation unwrapping, honestly.
-Billy