Re: Class Notes for a PL/I Course.

Am 06.10.2019 um 10:59 schrieb Giuseppe Vitillaro giuseppe-yuD/ahkh7LvrZ44/[email protected] 
[H390-MVS]:
>
>
> By the way reading around, I've found these references
>
> PL/I(F) 1966 IBM - New York Programming Center - Hursley (UK)
> BCPL 1967 Martin Richards - University of Cambridge
> C 1972 Dennis Ritchie - AT&T Bell Labs
>
> which drive what seems, to my eyes, a thread rooting
> these languaes to ALGOL (1958), through a strict relationship
> between them.
>
> That would explain why our PL/I(F) lack LEAVE/ITERATE.
>
> I bet we would find them in BCPL loops? Hence those constructs
> would have born between 1966 and 1972? Probably with BCPL in 1967?
>
> Anyone have time to add real history references to
> this too simple scheme?
>
>

If you don't yet know it (which I doubt, because it is well known in the 
community),
you should read this:

https://fermatslibrary.com/s/the-emperors-old-clothes

by C.A.R. Hoare

on page 82, he mentions PL/1.

Some things I heard about programming languages history:

IBM "invented" PL/1 together with the S/360 product line, and it was 
meant to
turn all other languages obsolete. Special regard was made to Fortran 
and COBOL;
these two languages dominated the scientific (Fortran) and the 
commercial (COBOL)
market at that time. ALGOL was a European language, defined in 1958 by 
Rutishauser
et. al. and was not very interesting in the USA at that time, but there 
were many ALGOL
compilers on European hardware. (Niklaus Wirth, BTW, did or supported a 
very early
port of ALGOL to the 360 machine, called ALGOL W, which was pretty 
successful).

It was very late when IBM decided that it was necessary to add
the ALGOL block concept (which was unique to ALGOL at that time) to PL/1.
Nobody today could imagine PL/1 without this, and all modern languages
of course have the ALGOL block concept (in this sense, COBOL and Fortran
are not modern, sorry). AFAIK, the first experimental versions of PL/1
(then called NPL) didn't have it. The name NPL (new programming language)
had to be dropped, because the acronym was already in use in the UK for
national physics laboratory.

C is, AFAIK, a direct offspring of BCPL; there were older languages called
A and B. I don't really know BCPL, but from what I heard, BCPL is an 
advanced
version of B, and, in contrast to C, it had no other types than int.

I'm very interested in Programming Languages History, but is it not 
always easy
to verify all the influences one language had on another language; it's all
about people, you know ... nobody knows what the language designers at
that time had in their mind, we only see the results today (if they didn't
write down their motivations).

Kind regards

Bernd


> .
>
>