Hi Winston,

Replying to a message of Winston Smith to Roger Scudder:

 WS> ForTran is the high level language version of the IBM 360 Assembly
 WS> language

This is incorrect.

The earliest commercially available FORTRAN compiler was for the IBM 650,
and was available some 7 years before the System/360 was released. The
FORTRAN language's syntax bears no resemblance whatsoever to S/360
assembler. Moreover, the S/360 assembler language offers semantic
possibilities that are not available in FORTRAN. [I have programmed
professionally in both these languages for some decades.]

 WS> (BAL for Basic Assembly Language? I forget what the
 WS> Assembler is called).

One sees both "BAL" and "ALC" for IBM mainframe
assembler. Neither is etymologically correct, but these acronyms are used
nonetheless.

The current assembler, for the S/390 and zSeries machines, has a program
name of ASMA90. Its predecessor, for 370/XA machines, was called IEV90 and
its predecessor, for S/370 machines, was called IFOX00. Between IFOX00 and
IEV90 there was an assembler from Stanford University's linear accelerator
lab., named ASMGASM, and I will leave it to the reader's imagination to
guess what JCL coding errors that caused.

 WS>  It copied all of the IBM data structures right
 WS> down to the Hollerith declarations for declaring string constants
 WS> that conform to the Hollerith tabulation machine census data.

Not so. The earliest FORTRAN compilers used the BCD interchange code to
hold "Hollerith" constants. The IBM S/360 used EBCDIC instead,
and FORTRAN Holleriths on that architecture were (and still are) stored in
EBCDIC. The Hollerith punch codes themselves were seldom used internally on
any machine, and since card readers are very, very rare these days that can
now be taken as "never".

 WS> ForTran is to IBM Assembler what C is to DEC Assembler Macro-11.

That's about right, actually. ... :-)

Regards

Dave


--- FleetStreet 1.25.1