On 7/01/19 2:52 PM, rbowman wrote:
On 01/04/2019 09:34 AM, Avi Gross wrote:
Although I used FORTRAN ages ago and it still seems to be in active
use, I am not clear on why the name FORMULA TRANSLATOR was chosen. I
do agree it does sound more like a computer language based on both the
sound and feel of FORTRAN as well as the expanded version.
It made sense at the time. I first learned FORTRAN in 1965 in
engineering school. At that time 'computer science' was in its infancy
and our everyday tool was a slide rule. The computer, an IBM System
360/30, was seen as another useful tool and engineers should learn to
translate their formulas into a form acceptable to it. You wrote your
efforts on coding forms, laboriously transferred those to punch cards,
and offered your deck up to the priests who fed it to the god visible
behind plate glass in his air conditioned lair.
Prior to FORTRAN, particularly in the pre-360 IBM mainframe world, the
only choice was Assembler - a 'language' which was merely a one-to-one
restatement of machine instructions in an acronym-like form. The issue
was that every machine type's instruction set was different and
consequently every Assembly language was different, ie there was no ONE
Assembler 'language'.
System/360 changed all that. (Brooks's book "The Mythical Man-Month" is
still a recommended text and a salutary tale) Now we had a series of
machines, at different sizes (the 360-30 was towards the bottom-end, or
one of the early sales - depending upon when in the time-line you look),
but as far as software was concerned, one machine behaved like the next.
(and IBM very much hoped that we would 'grow' and thus regularly need to
upgrade the processor - as well as adding peripherals... An IBM salesman
was not just for Christmas, he was for life!)
This practicality fuelled the (international) standards effort. It
became possible to have a (single understanding of) FORTRAN (others have
noted that there were in fact, implementational differences and matters
of scale), CODASYL went nuts with COBOL, and so-on (Ada anyone?). Also,
we had manufacturers attempting to impose, um, create a de-facto
'standard', eg IBM and PL/1 - nothing like the behaviors of today's
Googles, Microsofts, Apples, etc; of course...
Back to software, languages, and specifically Assembler: every
computation had to be broken down and coded as individual "low-level"
instructions:
load 2 into register-A
load 3 into register-B
add register-B into register-A
(note the lack of variable names, also pre-dating the ideas of
stack-architecture and modern/Intel CPUs!)
After Assembler, FORTRAN 'made all our dreams came true', because we
were able to write the likes of:
PI = 3.14159
AREA = PI * RADIUS * RADIUS
Can you see the step 'up' to the *formula* part!
Then, we would run our FORTRAN source-code through the *translator*
(later the generic term became "compiler"). It would translate our code
into assembler/machine code (varying by machine), saving us from the
more laborious and pedantic task of expressing ourselves at that level.
Oh the relief!
--
Regards =dn
--
https://mail.python.org/mailman/listinfo/python-list
