C=64 somewhere mid eighties. Taperecorder only a few days later. 5"25 Drive several years later. Had both MPS 801 and 803 printers.
First PC 386SX20 with 2 MB RAM, 40MB harddisk and a Trident 9000 VGA card and an Epson LQ 400 printer (December 1989 or 1990)
All except for the MPS803 purchased at Allkauf (later renamed to Real, and nowadays Kaufland) in Heinsberg, because back then electronics was cheaper in Germany than in the Netherlands.
Software
On C64 I started in Basic and was working on a font editor (for C=64s iirc 7x6 fonts). Displaying it was too slow in Basic, so I learned assembler to accelerate operations. Later I mostly worked entirely in assembler, but with a toplevel Basic program that mostly only did some calculations of parameters for the assembler routines that it called. Something that I still do in Delphi/FPC nowadays for my image processing routines.
On PC to learn QBasic I initially started with a simple graphical GUI (more something like you would make with Graph, not something like Lazarus). QBasic was too slow, and I got TP4 from a teacher and switched to it (months) and TP6 and BP7 a few months later again because I went to university and they had site licenses that also covered student use at home.
I dabbled what with inline assembler (as much example code from SWAG used it), but a good 1 1/2 year later (1994) I switched to Topspeed Modula2. The compiler was slightly more modern than TP and had an extra large model to avoid 64k BSS datasegment issues. I did a lot of work there and got more familiar with assembler. My biggest PC program from that time was a shareware cost analyser for Fido networks based on a logfile and nodelist scanner combined with the (telephone) telecom structure in the Netherlands. It mostly was used to measure how much Fido backbone costs a node had occured in a period, since those costs were shared.
I had maybe 5 licensees and some of them paid in old hardware, not money, but the primary benefit however was getting some more CS/IT savvy contacts.( I studied chemistry at the time)
For that work however, the 640k limit (and I was already using EMS) was a serious limitation and in 1997 I started looking out for a 32-bit Dos compiler. (While I moved to with win9x OSR1, I mostly only booted it for a few games and programs for school). I failed to get GPC working at all. In desperation I checked out DJGPP, but that was also a hit and miss, and example code was scarce and used unix derived libraries that were not a good fit for Dos in my opinion back then (and still).
So I ended up with FPC somewhere in april 98 I think. I ported some of the assembler code over, but not the cost analyser. Fido was declining, but the big killer was that most international and backbone connections went to internet, cutting out many of the reasons for cost sharing amongst backbone nodes. I kept submitting fixes for the dos and Linux RTL, and in end 1999 started to port FPC to FreeBSD because that was what my university account used. Somewhere in early 2000 I became core and had to figure out CVS (aargh)
Recent
