This is the archived version of Roland Weigelt's weblog that ran from 2003 to 2023 at weblogs.asp.net

Archives

Archives / 2023 / June
  • 40 Years of Computing (Part 1: The First 20 Years)

    This year, I have hit a major milestone: I have been using computers for 40 years. That makes me feel incredibly old. But, at the same time, I am also grateful because I have been able to witness the exciting home computer revolution firsthand.

    Here’s a timeline of the first 20 years:

    • 1983
      • I write my first lines of code on my cousin’s Sinclair ZX81 (with 1 kilobyte of RAM!)
      • A few weeks later, my parents buy a TRS-80 Color Computer at a local “Tandy” store. They get promptly scolded by our neighbor, who asks what kind of parent would buy their kid a computer.
    • 1984
      • Choosing a Color Computer turns out to be both good and bad. The reasonably good version of Microsoft Basic encourages programming. On the other hand, the computer remains virtually unknown here in Germany.
      • With few available resources and no one around me who actually knows how to write software, I struggle to program anything useful. That changes when I get my hands on imported US magazines (Rainbow, Hot CoCo). Tinkering with the listings of simple games to improve gameplay teaches me a lot.
      • Encouraged by my progress, my parents are open to buying a new computer for Christmas. The Commodore C64 is super popular, but the Schneider CPC 464 (a rebranded version of the Amstrad CPC) with its 80x25 character mode looks like a better alternative for doing more than just playing games.
    • 1985
      • My first larger program allows users to create simple animations that are converted into self-running BASIC programs. At a computer fair in nearby Cologne, I demo the program to other “computer kids” (back then, it was usual both at computer fairs and at department stores for people to meet and demo their stuff). I get approached by the co-owner of a software company. Long story short: they buy the rights, cripple/slow down the program with an incredibly bad copy protection and try to screw me out of the royalties for the measly 90 copies sold. A friend of a friend of my father is a lawyer and writes a nice letter that leads to a happy ending (for me).
      • Towards the end of the year, it becomes clear that I need to learn Z80 assembly language. Somebody gives me the copy of a copy of a copy of the CPC firmware manual. The low-level API turns out to be wonderful even by today’s standards.
    • 1986
      • I publish my first listings in computer magazines. These are the golden years of home computing, with actual money to be made for everyone with an idea and a bit of effort.
      • My school buys two IBM PCs. This is where I see Flight Simulator for the first time.
      • A young teacher with little programming knowledge, tasked to teaching BASIC in a weekly meeting after school, asks me for help.
    • 1987
      • At school, I make my first steps in Turbo Pascal 3. The speed and the ease of use of the IDE is impressive. The compiler is done by some Danish guy called Anders Hejlsberg, who incidentally would turn out to have a recurring impact on my life.
      • At home, I develop a Sidekick clone called “Sideworx” that requires a memory expansion card. Computer magazines ignore it and without access to a BBS or the Internet to spread the word, the software has an impressive total number of four users.
    • 1988
      • I develop my most ambitious 8-bit application, a graphics program called Artworx. It features vector text, undo, pattern fill and a superfast circle routine, all in 42 kilobytes of available RAM. Artworx is “the listing of the month”, the original article is available as a scan (https://archive.org/details/045-amstrad-pc-international-1988-11-ocr, page 30).
      • As a reward for my successful completion of school, my parents buy me an IBM XT no-name clone. Unfortunately, it does not see much use after I start my mandatory service in the German army.
    • 1989
      • After finishing the military service, I start studying physics at Bonn University. In hindsight, choosing physics over computer science had its pros and cons.
      • I start learning C using Turbo C 2.0. Coming from assembly language, the concept of pointers feels familiar.
    • 1990
      • Turbo C++ 1.0 comes out, introducing me to the world of object-oriented programming. One of my projects is a text editor with keyboard macro recording/playback.
    • 1991
      • At university, I start computer science as a minor subject. The programming part is a bit weird, forcing me to go back to Turbo Pascal after really enjoying C/C++.
      • For a group activity in programming class, I end up with some other guy who looks kind of bored. He is impressed when I fire up my text editor to use the keyboard macro feature on a source file for some repetitive editing task. It turns out he owns a software company, and I get hired on the spot.
    • 1992
      • During a four-week English language course at Jacksonville University, I get to play around with some Macs. The experience is nice, but I am not as blown away as I thought I would be.
    • 1993:
      • A small utility I write for the 4DOS command prompt called “4PRMPT” gains some popularity on the German part of the Fido BBS network. I promptly (pun intended) fall into the “second system syndrome” trap when I start a partial rewrite. This turns out to be just one of the many private projects that I start and abandon around that time. Not very productive, but a good learning experience.
      • I try out DOOM after a long download from the local BBS, and I am completely blown away. My peers at the university share the same reaction. Long nights of LAN parties involving carrying heavy computers around follow.
    • 1994
      • Low-level VGA programming of the “high-res” 640x480, 16 color mode using inline x86 assembly is one of the weirdest programming experiences in my life. Setting a pixel to a specific color involves managing groups of eight pixels in four bytes spread over four bit planes. Writing a proportional font output library involves a huge amount of bit manipulation. The library turns out nice, but aside from a GUI prototype, it does see much use.
    • 1995
      • I publish my freeware tool “R8Worx” for manipulating Flight Simulator 5.1 textures (.R8 files) and see it mentioned in various scenery projects.
      • At university, I start my diploma thesis, developing a C++ library for accessing electronics modules on bus systems like VMEbus or CAMAC. Choosing C++ instead of C is met with criticism, and I have to benchmark various scenarios involving direct, indirect and virtual function calls. Little do I know that this would come handy later.
    • 1996
      • A recommendation by my professor allows me to visit the CERN school of computing. Two awesome weeks of meeting students from all over Europe. This is where I try Java for the first time.
      • Back in Germany, my promise of extensibility of my C++ library is put to a test when I agree to support a new kind of electronics modules, just a few weeks before the deadline of my thesis. The software side works fine, but the hardware side turns out to be tricky. The research on the hardware bugs becomes part of my thesis.
      • Meanwhile, the local basketball team Telekom Baskets Bonn starts their first season in the highest German basketball league. I leave a comment on the website suggesting the use of sound effects at certain situations. One thing leads to another, and I end up writing sound software and supporting the DJ for the next ten years (more on what happens next in part 2).
    • 1997
      • Another professor, who agreed to give a second opinion on the thesis, tips off a local company called Comma Soft that there’s an interesting student (me). At the job interview, I get asked whether I have any idea how a C++ compiler implements virtual functions (remember this is 1997). I answer that I do not know for sure, but that my C/C++ benchmarks (see above) show that indirect function calls using pointers and virtual function calls are similar in performance. My first day at work is eight days later.
    • 1998
      • I play around with early versions of Visual InterDev and the Windows Foundation Classes (WFC) of Visual J++. I like it, Sun does not.
    • 1999
      • At work, I develop a frontend framework and a JavaScript preprocessor. The time is not quite right yet, though. As manipulating the DOM does not work in all browsers yet, the code uses framesets, document.write() and repeated reloads in individual frames.
      • Netscape 4 makes my life miserable, with the infamous “Netscape resize bug” being just one among many bugs. Each new dot-version requires yet another workaround.
    • 2000
      • I experience Steve Jobs at JavaOne live on stage, being Steve Jobs. “We are working hard to make Mac the best Java delivery vehicle on the planet”. Let’s put it this way: strong speeches are not always followed up by long-term actions.
    • 2001
      • Our team at work decides to completely ditch JavaScript on the client side. Together with a colleague, I develop a model-driven application framework using C++ and XSL that runs on the server without any script on the client side. It uses only the most basic CSS features and table layout involving transparent single-pixel GIFs. It turns out to be rock-solid; some solutions using this technology run for more than ten, sometimes 15 years without breaking changes after browser updates.
    • 2002:
      • .NET 1.0 comes out. It is no coincidence that Windows Forms reminds me of WFC. I do like the philosophy, but UIs are so slow compared to Visual Basic 6 it is not even funny. Things would get better in later versions, but never reach the snappy startup of VB6 applications.

    To continued in “40 Years of Computing (Part 2: 20 Years of Blogging)