30 Sep
2014
30 Sep
'14
9:45 a.m.
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-
bounces at classiccmp.org] On Behalf Of Chuck Guzis
Sent: 30 September 2014 17:20
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: IBM1629 & "C" was RE: Who is the world's oldest working
programmer?
On 09/30/2014 02:02 AM, Dave G4UGM wrote:
I learnt a lot from coding Fortran II on the 1620 but I never looked at the internals. I
believe the IBM1401 and Honeywell H200 series had similar structures with extra bits in
the character to delimit fields and records in the store. I worked in a shop with a H3200
and I remember the assembler coders having issues with switching from such a structure to
the Honeywell L66/10 we bought to replace it. When I moved on to NERC who also had a L66 I
learnt "B" as we made extensive use of "B" for systems programming...
... and before any one asks I don't think even "B" would work as
"B" expects a "word" structured machine like the IBM/360 or Honeywell
H3200...
Dave
G4UGM
The problem with the IBM1620 is that it has
extra bits in the
characters that can't be used to store data. This mean that
implementing the "C" language in way which allows it to perform
meaningful work possibly on existing data would be challenging. I
can't see any point in implementing any language just to show it can
be done...
I wouldn't think that anyone would really want to see a C for a 1620 any more
than one would want to see a C for an LGP30. But it's an interesting mental
exercise that reminds us that what was considered to be a "computer" can
be pretty different from what we think today. In some ways, that's too bad,
as it limits our thought processes.
C for a dataflow machine, anyone?
--Chuck