Wow, great work, Mark. Thank you!
I found the commentaries on the triangle plotter and horizontal line routines very helpful - I'd always wondered about those. I didn't realise people managed to pack so much logic into single ARM instructions productively, but Braben seems to do it with every single line of code - e.g. lines like "LDR R6, [R10, R9, LSL #2] ", I'd have probably naively coded in multiple ARM statements. That must be why my old game-programming attempts always ran so slow!
I would love to go though your commentary and port the whole Lander game into a high-level language like Python or even MatrixBrandy basic. It would make the whole Lander logic very accessible. Then also the game would run in higher resolution and frame rate than was possible on the original ARM1. I guess in Python the landscape mesh could be built as a single Numpy array, and the Fourier-series method to generate the landscape could probably be written concisely in just a few Numpy operations. So the whole volume of code would drop hugely. It would be fun to do (But I don't have time right now. I wonder if anyone would be interested in open-sourcing that project?)
I found the commentaries on the triangle plotter and horizontal line routines very helpful - I'd always wondered about those. I didn't realise people managed to pack so much logic into single ARM instructions productively, but Braben seems to do it with every single line of code - e.g. lines like "LDR R6, [R10, R9, LSL #2] ", I'd have probably naively coded in multiple ARM statements. That must be why my old game-programming attempts always ran so slow!
I would love to go though your commentary and port the whole Lander game into a high-level language like Python or even MatrixBrandy basic. It would make the whole Lander logic very accessible. Then also the game would run in higher resolution and frame rate than was possible on the original ARM1. I guess in Python the landscape mesh could be built as a single Numpy array, and the Fourier-series method to generate the landscape could probably be written concisely in just a few Numpy operations. So the whole volume of code would drop hugely. It would be fun to do (But I don't have time right now. I wonder if anyone would be interested in open-sourcing that project?)
Statistics: Posted by mike12f — Sun Feb 25, 2024 10:42 am