[fpc-other] [fpc-pascal] Quick Modern Object Pascal Introduction, for Programmers

[brian]

On Mon, 20 Jun 2016 18:53:12 +0000, Mark Morgan Lloyd wrote:

>Luca Olivetti wrote:
>> El 20/06/16 a les 11:44, Mark Morgan Lloyd ha escrit:
>> 
>>> A private online service I use has somebody who still bears the scars of
>>> Olivetti attempting to write an operating system in (strict, unexpanded)
>>> Pascal.
>> 
>> Oh, the memories (and, no, I'm not related). My first job was in one of 
>> Olivetti's factories where they used such a system for production 
>> control (they ate their own dog food).
>> I was programming the PLC (well, part of it) and it had to keep in 
>> memory two hours worth of planning since that was the time the computer 
>> needed to reboot when it crashed ;-)
>> 
>> I doubt it was plain pascal though: according to this page
>> 
>> http://www.storiaolivetti.it/percorso.asp?idPercorso=564
>> 
>> it was concurrent pascal
>> 
>> https://en.wikipedia.org/wiki/Concurrent_Pascal
>> 
>> (even if I'm not sure the wikipedia page refers to the same thing).
>
>I think he said it was mid-70s, but I got the impression that it was 
>very much ISO-standard Pascal and it didn't have facilities such as 
>assembler interfacing. On the other hand I think Concurrent Pascal was 
>one of Brinch Hansen's earlier attempts, and like Pascal itself it might 
>have taken a couple of iterations to get right.
>
>The bottom line was that it was an unmitigated disaster :-)

There was more than one dog's breakfast of a Pascal compiler around in
those days, My own favourite candidate was Whitesmith's Pascal for the
PDP-11. Debugging had to be done in octal from Macro-11 listings, and
Whitesmith's produced some memorably large and inefficient code. I
have vague memories (35 years later!) of successive increment and
decrement instructions *on the same register*. When you're trying to
shoehorn everything into 32KB, albeit with overlays, that sort of
thing MATTERS. :( 

It was only after monitoring the compilation process via RMD and then
doing a bit of RTFM that we realised what was happening - Whitesmith's
Pascal, at least of that era, was a translator rather than a compiler.
It was translating the Pascal code to C and then invoking Whitesmith's
C compiler. 

We played with it for a few days, then moved on to Oregon (OMSI)
Pascal and never went back. I believe OMSI eventually became the basis
of RSX-Pascal, which didn't exist at the point we were looking at
compilers. 

Brian.