Development, round II

From: Peter Prymmer <pvhp_at_forte.com>
Date: Tue Jan 27 00:24:14 1998

classiccmp_at_u.washington.edu
Subj: Re: Development, round II

Max Eskin wrote:

>I just picked up a book on Macintosh Think C (MS Press, 50c, I didn't
>bother getting Macsbug and others, also 50c each). For one thing, does
>anyone have an extra/unvalued license copy of THINK C, version 2.1-5.0?
>Also, what was the first programming language (I mean not binary or
>assembly)?

According to Goldstine in "The Computer: from Pascal to von Neumann" the
first working modern stored program was a sorting routine that John von Neumann
wrote in roughly 1943 - and had working on the ENIAC/EDVAC by 1946/47, but
it was in machine code (binary).

An Assembler was available on the Princeton U/Institute for Advanced Study
EDSAC by the Fall of 1949. Herman Goldstine and John von Neumann wrote a
programming manual for it by 1951.

Short-code (which would by today's standards be classified as a type of
assembler) was available for UNIVAC machines by October 1952 thanks to
Logan, Schmit, and Tonik.

Heinz Rutishauser of the ETH in Zurich described the world's first compiler
in a preprint issued by ETH in 1952 (based in part on work that Konrad Zuse
had published in 1948/49).

Grace Hopper (who had started out working with H. Aiken at Harvard) developed
A0 then A1 and published results in the ACM Proceedings by 1952. By 1955 she
released A2 - which was popular on UNIVAC computers. She went on the become
instrumental in the development of COBOL.

FLOW-MATIC and MATH-MATIC were also available on Sperry computers in the
early(?) fifties. Remington Rand then developed a language called UNICODE by
1957/58 for use on UNIVAC 1103A and 1105 machines.

John Backus (et al.) at IBM developed: 1] Speedcoding in 1953 for the 701
2] FORTRAN starting in 1954 (which took 18 person years to develop, and
was released as Fortran I in 1956/57) 3] served on the international committee
that developed Algol (starting in 1959 but continuing through the 60s).

>Another thing: CP/M was run on just about everything, usually with
>about 64K ram. How is it that MS-DOS blew up to about 384K? What
>did they put in there?

A DOS kernel is distinct from a DOS installation. The DOS kernel is tiny
even for fairly recent versions. On a Dell boot disk for
MS-DOS (I think 6.22 and thus not classic) I see the following file sizes:

  IO.SYS 40,774
  MSDOS.SYS 38,138
  COMMAND.COM 54,645

and on a bootable PC-DOS 7.0 (again not classic) partition I have:

  IBMIO.COM 40,614
  IBMDOS.COM 37,066
  COMMAND.COM 52,956

and in either case the configuartion files AUTOEXEC.BAT and CONFIG.SYS are
each less than 1 kB. The big user interface difference between these and CP/M
is the full hierarchical file system. Of course they also come with loads of
bloatware - but some of that is quite fun. e.g. PC-DOS can be optionally
installed with Rexx and I chose that option. I also have a couple of
different DPMI's available including the one for DJGPP.


Peter Prymmer
Received on Tue Jan 27 1998 - 00:24:14 GMT