On Wed, 06 May 2009 08:54:14 -0400, f...@thefsb.org (Tom Worster) wrote:

>clancy, i can't argue with you. my desired usage of break is really just a
>cover-up for a goto. i know.
>it makes no logical sense but i think i'd sooner adopt oop than gotos. my
>mom taught me to program back in the late 70s and early 80s. she was an old
>hand. when FORTRAN 4 came out she thought it was the bees knees. when Z80
>micros with MS-BASIC came out, she thought they were cute. when turbo pascal
>came out on CP/M, she was impressed and taught me to quit using gotos.
>so while it makes no logical sense, perhaps you can see that it makes
>emotional sense.

I can understand your reluctance to disregard your mother's advice, but 
unfortunately she
had been brainwashed to accept the dogma of the day. I could never understand 
the hysteria
relating to GOTO. Certainly it could be abused, as I knew to my cost, but it is 
clear and
explicit, whereas both break and exception are disguised GOTO's with 
ill-defined targets.

I started programming in 1967, in Fortran. There were only the most basic 
manuals, and CSIRO (for whom I worked) had a little computer (a CDC3200, with 
32K of 24
bit words, and costing only $500,000) in each capital city, and a big computer 
(a CDC3600,
with 64K of 48 bit words, and costing $2 million) in Canberra. Our local 
computer was at
Clayton, and I worked at Highett, so a courier collected our punch cards twice 
a day and
took them to the local computer, then brought back the results of the previous 
run, giving
effectively one and a half runs a day.

When I got ambitious, and needed to use the big computer, my cards were put on 
to mag tape
at Clayton, and flown to Canberra, where they were run through the 3600 
overnight, and the
results written back to mag tape. Next morning the tapes were flown back to 
driven to Clayton, run through the 3200 to produce listings, and these were 
then delivered
back to Highett. The flights were often delayed by fog in Canberra, and on 
average we got
three runs a week.

Programming was in its infancy, and the idea of using a stack to handle 
subroutines had
not been introduced (at least by CDC). The Fortran provided an assigned GOTO, 
which really
was the perfect instruction for writing 'write only' code. It also permitted 
you to jump
indiscriminately into, or out of, loops and subroutines, and it was probably 
abuse of
these options which gave the GOTO its bad name. 

I was developing a program for analysing linear electronic circuits, and 
developed my own interpreted language. The program was very simple; it 
consisted of a loop
containing three assigned GOTO's:

start:  assign begin to switch_one
        assign  ........

next:   read the next character
        if it's a number, GOTO switch_one
        if it's a punctuation mark, GOTO switch_two
        GOTO switch_three

        GOTO next

I left CSIRO in 1973, and did not have access to a big computer until about 
1983. By this
time the assigned GOTO had long since vanished, and I had great difficulty 
my original logic, until I unrolled the inner loop into a logical progression 
through the
possible inputs.

For the next 20 years most of my programming was in 8x86 MASM. This also had 
the GOTO, and
I was able to write extremely complex programs, despite its inherent verbosity, 
developing subroutines to handle all the basic procedures, and using GOTO's to 
define the
control structure.

PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php

Reply via email to