Mr Hermannsfeldt writes:

<begin extract>
Now, it is true that DFP helps with some of those problems, but when
programming in a high-level language one generally doesn't know what
kind of floating point will be used. Some, like HFP, give a truncated
quotient on divide (except on the 360/91), others a rounded result. If
you want identical results on all systems, you have to be very careful
with rounding modes. (Even if you know its DFP, you might not be able
to set the rounding mode.)
</end extract>

I am happy with his concession that "DFP helps with some of these
problems", but I find the rest of this passage  puzzling.  I cannot
imagine not knowing what kind of floating-point is used or selectable
by a statement-level procedural language (SLPL) implementation I am
using.

I use PL/I a lot, particularly for throwaway, investigational
routines; and for my purposes I most of the time choose to have it use
BFP.  Rounding modes I specify in trivial, macro-generated
assembly-language subroutines.

SLPLs are convenient, but they always represent someone else's (not
always congenial) world view.  When I can, I avoid C and its dialects
because I find their world view uncongenial, and in general I use
assembly language to modify the behavior of an SLPL when I find that
behavior unsatisfactory.

I have no objection to the teaching of a little programming to
demystify computing (in much the way that a little physics is taught
to secondary-school students).  That said, people who cannot tame an
SLPL, make it do their wills, are not professional programmers in any
meaningful sense.  It is not clear to me that they any longer have
useful work to do.  (There was a time when one needed to be able to
program a little in order to use computers at all, but that time is
long past.)

John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to