Hi Paul,
> I did have a play with BCPL that it was developed from though. ... I
> think B was then developed into C, etc.
Yes, BCPL was by Martin Richards at Cambridge University and was a
stripped-down version of CPL, the B meaning basic. That meant it was
easier to write a compiler for it, especially on small machines. There
was also a BCPL ROM for the Acorn BBC home computer, not surprisingly
since it also came from Cambridge.
Ken Thompson, creator of Unix, stripped it down further, to B, when he
wanted an even smaller language. It wasn't compiled to machine code,
but threaded code. Dennis Ritchie took over from him, creating New B
and then C, and changing the compiler to produce machine code along the
way.
A very good and popular book on writing compilers from my youth used
BCPL as the implementation language because, being simple and only
having one type, the machine word, the language got out of the way of
showing the compilation techniques. The author, Richard Bornat, has put
a version online now. The original book I have was produced in a
fixed-width font with underlining and over-printing for bold. :-)
http://www.eis.mdx.ac.uk/staffpages/r_bornat/#compilerbook
I can recall marvelling at how a C compiler could take source code and
produce machine code from it. I could write both but didn't know the
magic of how a compiler went from one to the other. Bornat's book
showed how to me and many others.
Cheers,
Ralph.
--
Next meeting: Crown Hotel, Blandford Forum, Tuesday 2010-11-02 20:00
Meets, Mailing list, IRC, LinkedIn, ... http://dorset.lug.org.uk/
How to Report Bugs Effectively: http://goo.gl/4Xue
