A history of making.

My first computing experiences were not part of a plan to become an engineer. I wanted to play games on a computer in my father’s office, then wanted to understand the commands that launched them. Books about DOS, BASIC, Pascal, and C followed. The more I learned, the more interesting it became to make the machine do something of my own.

That pattern eventually produced the original version of this website: a sprawling place for programs, research, friends, jokes, and a CV, all sitting over a pale circuit-board texture. I still like that it was unmistakably personal. The design here is more restrained, but I do not want to polish away the curiosity that made the old site worth keeping.

JulTools was my first contest entry: a DOS file manager meant to make copying, moving, and deleting files less error-prone. It won first place in a 1997 programming contest in Stara Zagora. The program was hardly elegant, but it did the work the brief asked for. That mattered to a young programmer who was discovering that a useful tool could be judged by someone other than its author.

The next year I built JulDraw for technical drawings. I spent more energy on its mouse-driven interface, including navigation inspired by AutoCAD, and less on some of the mathematical conventions the contest expected. It placed fifth. I still think of it as an early lesson that product quality has several dimensions: an interface can improve while the overall solution misses an important part of the problem.

FlashText, made with Smilen Kouzmanov, pushed a text editor beyond ordinary single-range selection. It could select multiple or rectangular regions and combine selections in ways that made complex editing easier. VT World was a different kind of contest entry: Peter Velichkov and I built a presentation of our home town in about two weeks in 2001. I worked on the dynamic content in Perl. Between them, these projects taught me that the interesting challenge is often not the algorithm alone, but what it lets another person express or discover.

Some projects existed simply because I wanted to learn. Flamer was a multipart graphics demo written for DOS in 1997. JulTris began as a text-mode Tetris experiment, then became a Windows game using DirectX. I remember these less for the code than for the pleasure of figuring out what the machine could do next.

Others came from annoyances close to home. A CD Cataloger kept track of files across physical discs and helped find the right one later. A text converter, which grew with help from Smilen Kouzmanov, bridged DOS and Windows Cyrillic encodings. Lihvi calculated interest for my grandfathers, replacing a slow, error-prone job they had been doing by hand. None needed a product launch to be worth making.

There was more formal work in the same spirit. During a placement at Siemens I wrote my bachelor’s thesis on a common interface for linear-programming solvers, including the data structures needed beneath it. In 2006 I also returned to my old mathematics and science high school in Veliko Tarnovo to talk with students about programming competitions. Those conversations have stayed with me: sharing what you have learned can be as energizing as building the next thing yourself.

The old project list also had an “Others” section, which may be the most honest part of it. Netmaker was a rough first attempt at Windows programming with MFC and graph algorithms. A computer won in a 2005 contest became a home server for a personal site. The server’s name, Grizach, came from an in-joke with Peter Velichkov about chewing through problems like rodents.

Peter and I also patched an NVIDIA Linux driver to work around conflicts with then-current 2.6-series kernels. I helped improve the Overclockers Online site after years in its forum, and made plugins for LCDHype before starting an unfinished Linux port, LCDHype4L. These were not pieces of a tidy career narrative. They were ways to help a community, learn a system from the inside, and make friends along the way.

I still value that kind of work. It begins with a concrete problem and a person who cares about the answer. Sometimes the result is a lasting product; sometimes it is a patch, a conversation, or a half-finished experiment that teaches you something useful years later.