The Communist Apple II, and a Story of Not Knowing What You Were Testing for 14 Years

Bringing back reverse engineering as Friday's topic, connecting how building technology from the front and understanding it from the back are ultimately the same problem.

The story of Bulgaria's Правец(Pravetz) computers is the starting point. The IMKO-1 was essentially a direct clone of the Apple II, matching it almost exactly down to the 6502 CPU, ROM, and schematics. What was added instead was localization such as Cyrillic characters, a metal case, and a heavy power supply.

The author views this not as simple cloning but as technology diffusion and learning under constrained conditions. Bulgaria spread computers to schools and research institutions this way, and says a generation of engineers learned computing by reading and reconstructing schematics.

Next comes the ISCAS-85 benchmarks. These circuits were standard test netlists released in 1985, but researchers ran tests on them for 14 years without knowing what the circuits actually computed.

That gap was filled in 1999 by Mark Hansen, Hakan Yalcin, and John P. Hayes. By decomposing each netlist into RTL block units and recovering their functions, they revealed the following actual purposes:

• c432: 27-channel interrupt controller
• c880: 8-bit ALU
• c6288: 16×16 multiplier
• c7552: 32-bit adder/comparator
• c499 / c1355: 32-bit single-error-correcting circuits

The key point is that this work is not simply historical restoration, but shows that diagnosis, synthesis, and reverse engineering are different directions of the same formal problem. Recovering a function by looking at a circuit, and building a circuit that satisfies a desired function, ultimately share the same ∃∀ structure.

Finally, the author places his own work on LyDiA and qbf-designer on this same continuum. Understanding past circuits and benchmarks once again reveals the value of 'explainable understanding' that today's EDA and AI can easily miss.