Re: About those parenthesis....

Majorinc wrote:
> I like parenthesis and prefix from beginning. However, I think
> that usual f(x,y,z) would be better choice than (f x y z).

The commas don't serve any purpose, other than to add clutter. So a
better notation than f(x, y, z) is quite simply f(x y z).

I have programmed with Lisp-like semantics using the f(x,y,z) syntax.
It's ugly. The comas really are a problem. There is no satisfactory way
to format them. The syntax does not scale to argument lengths required
in Lisp programming, where long expressions have to be split across
lines.

The commas have to go. At that point, it makes sense to "castle" the
function by exchanging it with the parenthesis.

Hanging on the outside of the parenthesis, the function isn't
syntactically encapsulated. There is an ambiguity there. A
left-to-right scan, having seen only the identifier f, cannot decide
whether f is a plain term, or an operation with arguments. The scan
must continue to the next token before that can be decided. Until then,
the scanner is in a state that is a combination of two states: ``I am
parsing either a simple term, or a the beginning of a compound''. If
you put the parenthesis first, it's clear. You read a token. It's a
parenthesis: great, we have a compound. It's an identifier, so we have
a term.

The f(x y z) notation has no nice way to denote the empty list, because
the token f is needed for syntax. The notation () is a syntax error
because it has no operator on the left.

The f(x y z) notation is also very poor when it comes to subexpressions
which do not represent an operation f being applied to arguments x y z.
It assumes certain semantics: namely that the left item is elevated
above the others. This is not the case when you have syntactic
abstraction.

Consider what happens to this:

(defclass c (b1 b2 b3) ())

Let's skip the fact that we don't know what to do with the empty list
of slot definitions (). In fact, let's invent the @() notation for
that, what the hell. So we get:

defclass (c b1 (b2 b3) @())

Do you see how wrong that is? (b1 b2 b3) is just a list of base
classes. Multiple inheritance from three ancestors. This is mangled
into b1 (b2 b3) which doesn't make sense at all. b1 isn't an operator,
it's just a base class like b2 and b3.

Even when the first symbol /is/ an operator, it's not more important
than the other two symbols.

When you write (+ 1 3) to denote 4, the 4 is a result of all three
symbols acting together. Each one contributes as much to the result as
the other two. The + does not deserve any more credit than the 3 for
that result. It's the combination of these three symbols in that order,
and the surrounding broader context, that gives rise to 4.

So, in summary, f(x, y, z) is an abomination that rightfully deserves
to be cast into the lake of fire.

.

Relevant Pages

  • Re: a^b^c ?
    ... there is the simpler notation => c. ... so it is not important put parenthesis ... for say the true i wrote a big copybook [quaderno] ...
    (alt.lang.asm)
  • why eta(958)?
    ... Can anyone clarify me why the eta' is named with a mass in parenthesis? ... I believe this notation is reserved for particles decaying strongly, ...
    (sci.physics.research)
  • Re: Is there a better/simpler way to filter blank lines?
    ... expression from surrounding syntax. ... surrounding syntax with the exception when there are already enclosing ... So parenthesis are tied to generator expression syntax. ...
    (comp.lang.python)
  • Re: Confusion with bindings - scheme newbie
    ... the syntax of LET is ... a LET that has only one binding looks like ... [(sos y z)] ... Now thats a lot of parenthesis to deal with. ...
    (comp.lang.scheme)
  • Re: ksh function names
    ... The syntax with or without parenthesis makes it work differently, ... >> myfunc() ... but in recent ksh versions it returns the global script ...
    (AIX-L)