BCPL, a predecessor of 'C'

http://www.lysator.liu.se/c/clive-on-bcpl.html

by jacquesm • 17 years ago

18 17 17 years ago

lysator.liu.se

tjr• 17 years ago

Interesting that BCPL included both one-line and multi-line comments, but C did not.

jacquesm5• 17 years ago

I used the BBC micro version by Acornsoft in the gray past, it made the transition to C a lot easier when I finally got my hands on a C compiler (on a PDP 11 bought for scrap...).

draegtun3• 17 years ago

Yes Acorn/Acornsoft did do for some "off the track" languages.... I recall they did COMAL as well.

mahmud• 17 years ago

Kaz took a stab at it recently and found a few surprising features that didn't remain in C:

http://groups.google.com/group/comp.lang.lisp/browse_thread/...

In Summary:

I didn't say BCPL is better than its successors. Only that not everything in its successors is strictly an improvement over its BCPL counterpart.

jacquesm5• 17 years ago

I think the reason for a good bit of that is that BCPL was first stripped to produce 'B', then expanded again to produce 'C'.

ajb• 17 years ago

BCPL's predecessor, CPL, had even more features: things like type inference, closures, polymorphism. It was never completely implemented. The machines at the time were too small - 4kwords being a typical size. It would probably be fairly easy nowadays.

jlees• 17 years ago

Oh dear gods no.

I first encountered BCPL when I was at Cambridge University, the home of the language and its designer, Martin Richards.

Same. I was taught this during my undergrad (in the ancient times of 2002-3ish). A course on "Comparative Programming Languages" ended up as 7 lectures on BCPL and 1 on everything else (Lisp, Smalltalk, stuff like that. I think.) It was terrible.

I now have an in-built knee-jerk reaction to hate BCPL. Especially since he (Martin Richards, who should never be allowed to lecture, ever) spent at least one lecture teaching us how to implement types in BCPL.

Yeah. Typed BCPL. Last time I checked, that was called "C".

jacquesm5• 17 years ago

That's absolutely amazing, I thought BCPL was long dead.

Or rather, it seems that it is long dead but that this particular lecturer can't get off his hobby horse.

That's archaeology, not 'comparative programming languages'.

It's a bit like going somewhere to study French and getting a course in Latin instead.

RiderOfGiraffes2• 17 years ago

I liked BCPL, and still prefer it to C, but I taught it to myself, and never took lectures from Martin Richards.

Most of the computing done for my PhD was done in BCPL. I still have the code somewhere, but I can't convert it to C because it uses co-routines.

jacquesm5• 17 years ago

http://www.dekorte.com/projects/opensource/libcoroutine/

ajross• 17 years ago

While I'm sure you understand the distinction, I think a lot of people miss this:

When BCPL or B are said to be "untyped", it's not in the same sense as Lisp is dynamically typed. These are languages for word-oriented computers, without byte-addressable memory. Everything the CPU can do is on a single word of the same size, and memory is just an array of them. So it's untyped in the sense that in C you can do math on both integers and pointers, and convert between them.

What C added was originally byte orientation, "typing" was just the glue required to make that work (so that e.g. an int was 2 chars on a PDP-11). Only later on were first class arrays and structures added to the "type system", and likewise the space of "pointers to types".

russross• 17 years ago

I thought maybe it was just me. When I was a postgrad at Cambridge (2003-2006) I supervised for this course a few times, thinking that a course surveying a bunch of interesting languages would be fun. I seem to recall a bit of smalltalk thrown in, but it was mostly BCPL. I walked by Martin Richards' office and noticed a poster announcing a port of BCPL to .NET or some similar abomination.

It's time to let it go...

draegtun3• 17 years ago

Don't see any mention of O-code in that article?

O-Code Machine, a predecessor of JVM:

  http://en.wikipedia.org/wiki/O-code_machine

RiderOfGiraffes2• 17 years ago

Clickable ...

http://en.wikipedia.org/wiki/O-code_machine

jacquesm5• 17 years ago

Doesn't the JVM go all the way back to UCSD p-code and Forth ?

draegtun3• 17 years ago

And O-code looks like it also precedes them.

BCPL, a predecessor of 'C'

Related Stories