Do fair comparison

[lmartorella]

Here https://leanchess.github.io/ you mention the 1K ZX Chess, comparing the size of the program as a "score".
However, you should note that the unexpanded ZX81, for which the program was built, has really only 1K of RAM. This means that the whole application code, the CPU stack, the variables, and even the board screen displayed (as ASCII chars) should fit in such tiny space.

Then, to play fair, you should take in account the whole buffer_db usage, the x86 max stack size and even the RAM used by your video card to show the board. That's the only way you can compare against 1K ZX Chess :)

[dmitry-shechtman]

All valid points, but you're comparing apples and oranges. The hardware is quite different, but I'm (successfully) competing in the software department.

[lmartorella]

Sure, you are competing in a relatively new field, that is very similar to the 1k/4k/64k demoscene: minimize the bytecode size of an application (even using compression or similar techniques), but counting on the availability of modern hardware, and more important, memory. Today a x86 segment of 64Kb seems ridiculously small. But once, it was more that an average system could bring.

Making a chess application that actually runs in 1Kb is just another league. And... it is software department as well. Apples with apples, but smaller. Sort of adding push cs; pop ss; mov sp, 400h at the beginning of your software and see if it still works.

So, in the end, you can compare the size of ZX81 applications just for reference, but please remove references like "less than half the length of the famous 1K ZX Chess".
Oh well, unless you make it running in such low memory!