Back in 1983 I joined with several other people to start a company and build a neat office automation product and word processing product based upon the Intel 80186 chip which was brand new at the time. I worked my butt off 14 to 18 hours a day for 1 1/2 years building some awesome hardware. I even used my new Corona Data Systems dual floppy PC clone machine with the IBM MASM assembler to write full diagnostics and test code to exercise and validate the hardware of my awesome 80186 hardware creation.

Meanwhile one of the other original partners and two programmers he hired were to devise the operational software that was to make this machine function. They could have gone right to work making the product software using a PC and MASM as the cross assembly station. But guess what...that partner had the misguided idea that the product software had to be developed right on-product. So they (the three of them) set out and first wrote an assembler, and then an editor, and then a debug monitor, and then a crappy file manager for the 3.5 inch floppy (crappy because the directory was kept in a paper notebook and a file was a track on the disk). The debug monitor command to load the editor would be something like X 13 to execute the program from track 13. In the editor a command to edit a file was E 26 to edit the text from track 26. In the assembler the command to assemble a program was like A 38 to assemble the code at track 38. There was no linker as that was deemed "too hard".

We were into this company about 2 years before these clowns even were ready to begin the actual product code. And then the stuff they did make was done at an amazingly pokey pace due to their crappy home-spun tools. Now we were well into 1985 already and we could have even had XT clones sitting around with hard disks for doing software development....but now instead I watched a brilliant product concept fade away as the capital market place for startup financing went from good to bad and finally tanked in 1986 with the failure of the company.

My takeaway from that three year experience in futility was to not ever trust someone who says that they have to first make their own assembler before they can make the product program when there are viable commercial choices available for a couple hundred dollars that can run on hardware that is widely available.

Michael Karas