Architectural diversity - was Re: Pair of Twiggys
john at forecast.name
Fri Mar 17 10:16:20 CDT 2017
> On Mar 17, 2017, at 9:05 AM, Paul Koning via cctalk <cctalk at classiccmp.org> wrote:
>> On Mar 16, 2017, at 9:28 PM, ben via cctalk <cctalk at classiccmp.org> wrote:
>> But was FORTRAN that portable?
>> Other than the IBM 1130 I cannot think of a small computer
>> that had ample I/O and memory to run and compile FORTRAN. All the
>> other 16 bitters seem to more paper tape I/O.
>> I suspect 90% of all university computers ended up as IBM 360
>> systems. A few ended up with the VAX, but who knows what they
> I know of FORTRAN implementations for one's complement machines with word length of 24, 27, and 60 bits, decimal machines (IBM 1620), two's complement machines of 12, 16, 48 bit words, just to pick a few. FORTRAN implementations tended not to be all that demanding of resources: 4k words is a typical minimum.
> I think a lot of high level languages are quite portable. ALGOL is not as widely ported but not because it's inherently harder. PASCAL was ported to many different machines too. C is a bit of an anomaly because it's more like a high level assembly language, so it has portability limitations that many other high level languages don't run into.
I just released a new version of the CDC 1700 simulator for SIMH. This is a one’s complement, 16-bit machine and the Fortran compiler is now functional in 16KW of available space (a smaller version (12KW) was available but I don’t know if any copies survived). The source code for the compiler is available on Bitsavers - it’s written mostly in Fortran.
More information about the cctalk