Re: Ternary operator

[Hans Boldt <boldt@xxxxxxxxxx>]

Don wrote:
> Tom,
>
> good question, and we could also ask why people insist on developing new
> programming languages/environments versus improving what we already
> have...  Frankly, I've yet to see anything done in a current language that
> couldn't have been incorporated in good ole PL/I.....  :) :)
>
> Oh well........

Ahhh! "good ole PL/I"! A truly marvellous language, developed before 
computer scientists had any decent understanding of what's really needed 
in a programming language. But before Pascal, and before C, PL/I had 
everything already. Things like structured programming and user defined 
types. Its problem was that it had *too* much. The designerS of PL/I 
looked at the programming languages in common use in the mid 1960's. But 
instead of selecting a sufficient subset of available language features, 
they tossed all known language concepts into a bowl, stirred it up, and 
called the resulting conglomeration "PL/I". We've come a long way since 
then! ;-)

Back on topic, if you think coding:

    if a>b;
       res = foo;
    else;
       res = bar;
    endif;

is too much work, you could shorten it by assigning values to an array 
and indexing into it:

    arr(1) = foo;
    arr(2) = bar;
    res = arr(%int(a>b)+1);

Alternatively, you could define your own "ternary" procedure, and call 
it like:

    res = ternary(a>b: foo: bar);

But these alternatives are still not quite like the "ternary" operator 
of a couple of other languages since both result expressions are evaluated.

BTW, this reminds me of a long debate in the Python community, where no 
one could agree either on the syntax for a "ternary" operator, or the 
very need for one. And so a decision was made by the BDFL to not bother 
adding it to the language.

Cheers! Hans