Re: Precedence values for Reverse Polish

  ppaatt@aol.com (PPAATT) writes:
>> > a b = c 1 := c 2 := if
>
>> ... deferred ... lambda ... then your RPN evaluator can remain
>> extremely simple.
>
>Why isn't this approach more popular?


The straightforward implementation is more complex and less efficient
than the less elegant solutions; and if you want to make it efficient,
you get even more complexity.


Still, Smalltalk and Postscript use this approach.


>This Way removes the need for the infamously irregular (non-portable)
>"special forms" of Lisp, the "compile mode behaviour" of Forth, etc.?


In Lisp you would still need the "quote" special form, no?


And looking at Postscript's complicated rules for literal and
executable objects, this does not really appear to be an improvement
over Forth's interpret and compile state (which admittedly have their
share of problems).


>My current compiler project passes this kind of tree as the
>intermediate form between front and back end ... what is it about this
>approach that is going to bite me but just hasn't quite yet?


I think it would bite you if you tried to generalize and also allowed
procedures passed at run-time, not just procedures known at compile
time.


- anton
--
M. Anton Ertl Some things have to be seen to be believed
anton@mips.complang.tuwien.ac.at Most things have to be believed to be seen
http://www.complang.tuwien.ac.at/anton/home.html

Return to the comp.compilers page.
Search the comp.compilers archives again.