On Mon, 5 May 2014, John J. McDonough wrote:

> On Mon, 2014-05-05 at 23:01 +0100, smplx wrote:
>> On Sun, 4 May 2014, Byron Jeff wrote:
>
>>> I'm pretty sure if you examine the architecture and speeds of the machi=
nes
>>> these languages were originally developed for, you will fine that these
>>> supposed "short cuts" were necessary.
>
> C was developed a LONG time ago.  It was intended as a successor to B. B
> had no data types - the only thing it recognized was a machine word.
> The concept of data types was new, at least to AT&T. (Seems to me
> FORTRAN had data types at the time, but FORTRAN was for an entirely
> different purpose, probably considered as relevant as COBOL.)

Ah the good old days :-)

Yes I used BCPL briefly (the forerunner to B). As you say it only knew=20
about machine words. To play with text you had to pack and unpack vectors=20
of words.

It seemed strange using BCPL when I had previously used ALGOL 60 which=20
knew about types.

>
> I'm sure when coming up with the idea of typing data, simple rules would
> have been preferred to some complex rules, even if the simple rules
> misbehaved at the boundary cases.  The PDP-7 preferred for B was a very
> different machine than the PDP-11. I suspect the dramatically different
> architectures and the new concept of data types made the few
> pathological cases of little concern.
>
> One of the main interests in going beyond B was to take advantage of the
> ability to address a character, something not possible on the 7.  I
> imagine the subtleties of 16 versus 32 bit, signed versus unsigned were
> of little interest when the main feature was working with characters.
>
> If you have ever watched any of the early Unix videos, the AT&T folks
> were really geeked about the whole concept of using the computer to
> manipulate text.  That was clearly the focus, and back then, text was 7
> bit, so no need to get all hung up over signed and unsigned.

Text was 6 bits :-)

No seriously, back then text was whatever the hardware manufacturer wanted=
=20
it to be. I actually cut my teeth on a ICL 1900 mainframe. It had 24 bit=20
words and 6 bit characters (packed 4 to a word). I seem to remember the=20
DEC system 20 had variable length bit fields which meant you could have=20
any number of bits in your character upto the maximum word length.

Regards
Sergio Masci
--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.