Post #4265824

Crazy how I learned all about BNF and LR(k) and the big names of the compiler writing hegemony in school but I'm just recently learning about Pratt parsing. I mean, looking deeply at it, it's basically just trainyard shunting, but instead of declaring it a toy and leaving it at implementing a few calculator ops, it presses into the idea deeply enough to explore the whole gamut of unambiguous (deterministic) grammars, including those which aren't context free, and shows the approach to be far simpler and more extensible, producing much more attractive syntax trees, much better semantic integration, and and to be easily implemented in hardware. This feels like it's awakening something in me that went to sleep when I was still in college. I'm reminded of all the ideas I had that I would have tried to prove out if I hadn't been discouraged from even trying with prejudice. It reminds me again just how petty the academic fiefdoms really are. It's a shame how tired I am now