Early Assemblers

This was how the SAP assembler on the 704 worked. This assembler supplanted the NYAP assembler from IBM shortly after delivery of the first 704’s.

The input was a deck of cards, read thru the attached card reader. Most cards held the symbolic form of a machine instruction such as:

LAB    CLA T9,4   COMMENT

They were upper case only. Some cards held assembler directives as they do yet today.

An assembler listing was the printed output of the assembler program that would include one line of text for each input card. Included on each typical output line was:

The octal location to which the instruction was assigned
The instruction, in octal, as it would be placed in memory
The serial number of the card (not included on the card)
A copy of the cards contents.

About a year later the assembler was enhanced to produce a printed cross reference reporting the serial number of all of the lines on which a given symbol had appeared.

The other output of the assembly was the program punched in binary onto cards. There were 22 instructions on each card plus starting address and checksum. There was no relocatable output in the early days.

Just a few years later an improved assembler called FAP arrived. George Michael tells me it was by Dave Ferguson. It provided relocatable output and macros. Relocatable output and concomitant mods to Fortran made it easy to write assembler subroutines for Fortran.