john gilmore schrieb:
declare (pi value(3.14159_26535_89793_23846),
sqrt_pi value(sqrt(pi)) binary float(52) ;
What I find most interesting in this example:
will the sqrt(pi) function call be evaluated at compile time?
I hope so.
What I expect from an optimizing compiler, is, at least, that I don't
get punished if I use normal and pratical language elements, for example
structure = '';
in PL/1, to initialize a (big) structure, all elements with their proper
initialization values, depending on type.
But there were compiler versions which generated such horrible machine
code,
that in our shop we were told to define a STATIC INITed structure with
the same layout and
to do a physical copy out of this INIT structure.
Some of us refused to do so - argueing that it is the job of the
compiler to generate appropriate machine code.
My opinion: if the language allows to express things in an easy and
comfortable way,
the compiler must support this by generating efficient machine code.
Otherwise,
if it's not possible, it's better to leave the features out of the
language.
And: the user of the language should have an idea of the performance
implications
of the language elements he or she uses. I'm not sure, if this is always
the case - especially
with PL/1 and less experienced programmers.
Kind regards
Bernd
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
