Let me tell you about my first job. It was back in 1976. I was employed programming one of the first word processors. It ran on a computer called a Raytheon PTS-100, which was originally designed for airport login systems - it was had 64 or 128k of memory and up to 8 character terminals directly memory mapped into that memory (what Apple would now call 'integrated memory'). We had to program in assembler, because there was no HLL. The development process went like this. You wrote your code by hand on a pad. Then you went up to the Teletype room and typed it into the machine that ran the assembler, which was an IBM360 somewhere different. You printed it out on the teletype and hand checked to see that you got everything right. Then you entered the 'commit' command, and hoped that no error messages came clicking out of the teletype. If they did you had to go into a batch editor to make changes. Finally you'd get the assembly done OK. The next day you got a little stack of punched cards through the post. The you went to the PTS-100 in the computer room. Beside it was a stack of cards maybe half a metre long. In the middle of that was the cards you needed to replace, behind a card carrying a link/loader command. You took those cards out and replaced them with the new ones, and if what you did had changed the code size, hand punched a new link/loader card. Then you took the whole stack and put it in the card reader. It red in really quite quickly, and you used some commands on the front panel to dump it to the enormous 2.5MB hard disc. Now you could reboot the machine from the hard disc and see if your code did what it was supposed to. Anyhow, one day while doing this I tripped up while doing this and threw the cards all over the floor. I was sure I was going to get the sack. It was a Friday, so I sneaked into the office on Saturday to see if I could do anything. In the teletype room were a set of IBM/360 manuals, and I found this thing called JCL (Job Control Language). I worked out how it worked and produced a JCL program that I hoped would reconstruct the whole stack from the code base on the 360. I crossed my fingers and set it running. When I came in on Monday morning my boss called me over. On his desk was this big stack of cards that had come in through the post - he'd had to sign for them because it was such a big package. Of course, I had to fess up. WE went to the computer room, put the stack in the reader, read all the cards, dumped them to disc and rebooted the machine. It only worked. So instead of getting the sack I got promoted, because my JCL program revolutionised their whole development program. Now they could keep different versions, have candidate releases and all that whizzy stuff.