Re: References to "History of Compilers"

The moderator writes:
|> [I thought Autocoder was the assembler for the 1401. -John]


kanze@lts.sel.alcatel.de wrote:
: The 1401 had something called autocoder, but I don't think that it was
: really an assembler. My impression was that an assembler could handle
: some forms of symbolic addressing, whereas an autocoder couldn't.


My recollection is a bit clearer, as 1440 Autocoder was my first
programming language. The original assembler for the 14xx series was
SPS (Symbolic Programming System.) This, as its name implies, handled
symbolic addressing. It had a fixed-column format and some
symbol-length restrictions, but was otherwise remarkably similar to
the sort of dreck that passes for an assembler today, and ran on a 4K
1401 with no disk or tape :-) Autocoder was a "free-form" (fields
determined by syntax, not column number) assembler with a primitive
Macro facility. Well, actually the macro facility was primitive
compared to, say CDC COMPASS or DEC MACRO assembler, but a few
generations ahead of today's "macro assemblers" which are typically
macro-pre-processors bolted onto the front of the same old brain-dead
assemblers.


Autocoder allowed the use of macro packages such as IOCS
(originally written by Lockheed, as I recall, but distributed
by IBM back when computer vendors distributed useful source), which
abstracted the differences between the various 14xx machines and
the specifics of a particular installation, in such a way as to provide
a reasonably portable way to do I/O-intense programming. As I recall,
it allowed such things as designated error handlers (over-riding
the default "punt to the operator", now implemented as the "you're
screwed" dialogue box).


moderator again:


: [Gotta go check my refs. There were a bunch of Autocoder to Cobol
: translators sold in the early 60s. Must have been interesting. -John]


Using such packages as IOCS allowed much of the "hair" of
a robust application to be described in a well-defined (if not-
so-easily read or written) form. The translation of such code
to another language would be helped a great deal by such rigor.
The translation from IOCS file-specifications and autocoder
field-definitions to COBOL I/O and DATA DIVISION specs would have
been fairly straightforward. The PROCEDURE DIVISION might have been
more complex, but a language with "ALTER ... TO GO TO..." would
allow some, uh, interesting code generation :-)


Mike
| Mike Albaugh (albaugh@agames.com) Atari Games (now owned by Williams)
| (No connection to any company owned by the Tramiel family)
| 675 Sycamore Dr. Milpitas, CA 95035 voice: (408)434-1709
--

Return to the comp.compilers page.
Search the comp.compilers archives again.