Step 1: Get legal move generation, PGN/FEN parsing and a reasonable on disk storage system for games. Base this on stockfish data structures.
Step 2: Integrate into lakinwecker/delila
Step 3: Continue porting Stockfish and Scoutfish ideas/algorithms.
Step 4: Adapt to focus on teaching humans
This engine is not going to be focused on having the highest possible elo. Rather, it will be focused on helping to train players. The intent is to develop efficient identification and explanation of tactical and positional motifs. So that we can easily process a large set of games quickly, identifying various tactical opportunities as well as to map up each of those tactical opportunities to the appropriate motifs.