Re: MEX paper, draft 1.0, part #4 of 5

[cowan (John Cowan)]

la deiv. bo,n. cusku di'e:
> la djan. cau,n. cusku di'e:
> > In reverse Polish notation, the operator follows the operands, and there
> > are always exactly two operands.  No parentheses are required or permitted.
> 
> It is NOT the case that there are always exactly two operands.  What
> IS true is that the number of operands is known for each operator.

I should have clarified this.  I meant "In the Lojban implementation of
reverse Polish notation, all operators must have exactly two operands.
Those operators which do not in fact have two operands must be supplemented
with the null operator or the null operand, just as is used in Lojban infix
notation."

I will make this correction.
> > The operands of a reverse Polish operator can be any legal operand,
> > including parenthesized mekso that can contain any valid syntax, whether
> > more reverse Polish or something more conventional.
> 
> Perhaps, but RP is a parenthesis-free notation by design.  While you
> can certainly define things this way if you want to, I think it may
> just muddy the waters.  Probably this can only really be determined by
> experience.  BTW, note that the parenthesized material should *not* be
> RP, because if it were there would be no reason to parenthesize it.

If the parenthesized material were, in fact, RP, it would be necessary to
insert another RP marker.  The point is that one kind of operand is a
parenthesized expression, and this is true no matter which operator syntax
we are using:  infix, Polish, or RP.


-- 
cowan@snark.thyrsus.com		...!uunet!cbmvax!snark!cowan
		e'osai ko sarji la lojban