Back in 1980-1985, I wrote some software for employment agencies; it was an early database back when no one knew what a database was. It was originally written to run on an Apple ][ with dual floppy drives, and had lots of optimisation to minimise the amount of space used on disk. Eventually we moved to am Apple ///, and then on to a PC. Eventually we moved to SCO Xenix to gain the ability to have multiple users accessing the database; Xenix at the time was running on an 80286. This was at a time when Microsoft said that you couldn’t multitask or have multiple users on those systems, we had some customers with up to 32 terminals all running on the same system. Once we had moved to Xenix, we sold a number of systems to some fairly large employment agencies and even one insurance company. When SCO stopped selling Xenix and started selling Unix, the software ran on SCO Unix without any problems or changes needed. It also ran on PC-DOS & MS-DOS.
We were fairly successful for a while, and were in business for about 10 years, until 1990. By then much larger firms were moving into the business with resources we weren’t able to match.
In 1999 I updated the software to take the Y2K problem into account, and the existing customers continued using the software without any problems; the only calls I was getting was for minor enhancements.
Over the years our customers either went out of business, were taken over, or in a few cases, eventually moved to a new, modern software package. But I still had two customers who were using the software until this year! Because I’ve provided support over the years I’ve kept one development system in storage, an old 80486 running both SCO Xenix and SCO Unix.
Finally, however, I received word that my last customer is no longer using the software. So I decided to capture the data from the Xenix system and move on. Boy, what a trip down memory lane; Xenix didn’t have network software initially, and I never purchased the networking software for it, so here I had a computer, with 4 meg of memory and a 200 meg hard disk, which I have no real access to. The disk was so old that modern motherboards weren’t even able to recognize it, and Xenix is so old that it cannot access any modern hard disk due to the size. So I was finally reduced to setting up a null-modem cable between the old Xenix system and a spare Linux box, running minicom on Linux and transferring the data over the serial cable…..at a baud rate of 38400, which equates to about 3000 characters/second after error checking is done! It’s in progress right now, and will take a few days before I’ve transferred everything I need.
My software package was called: The Matcher, and was written initially in UCSD Pascal on the Apple, then ported to Microsoft Pascal on the PC. MS Pascal is an ISO compliant compiler, there were a number of changes I needed to make, but overall the porting process took less than a week (if my memory is correct, it was a looonng time ago!). The advantage to the MS compiler was that it compiled the exact same code on DOS and Xenix, without any syntax changes. I’m not a fan of Microsoft, and the compiler was not too well documented, but it worked perfectly!
I may try and compile it with Free Pascal, just to see if I can.
This page is a tribute to a software package that was in use for 30 years! Yes, I am tooting my own horn, if you will, and yes, it was running on an ancient OS and computer; but it ran, and was used productively for all this time. Say what you will, but before you do, show me another software package which can say the same. There aren’t many.