The very first idea for enterprise integration was to have a big central computer in a central processing location. For me, my first exposure was repairing a CDC 3200 system, channel controllers, tape drives and specialized device controllers to extract mission performance (roughly telemetry) and ELINT records (like asynch messages) from tapes recorded in flight. I learned FORTRAN and JCL on mainframes using batch processing and punched cards.
Time Sharing, Multitasking and Multiprocessing:
An early wave of integration technology was to have many terminals connected to the same central processor so folks would not have to enter punched cards or process in batches. My first exposure was a Gould/SEL 32/55 on which I also learned to program in BASIC. This minicomputer used microcoded bit-slice technology.
Embedded Microcontrollers and Microcomputers: At first microprocessors were thought of as integration lools to add some limited local intelligence to printers, terminals and other equipment. My first exposure was in maintenance of the AN/ALR-46 airborn warning reciever which used an integral real-time processor. I remember debugging (and eventually building my own with a soldering iron and chips) z-80 CP/M computers. I also developed embedded computer boards and programmed others. At one time I was fluent in 8080, 8048, 8051, 8096, 2650, 6502, 6800, 6809, 68000, 8086, 7800, z8000 assembly language.
Modems: At one time much of integration involved modems to remote locations. My first exposure involved rewriting modem-720 in z80 assembler and causing my z80 system and H19 terminal to connect to the campus network at Penn State. I later designed my own modem board once. I also spent a large chunk of time on asynchronous and synchronous data transfer and messaging over modems, resulting in a patent at CDC.
"If you do anything long enough, you eventually get good at it." MK