TODO.md

TODO

• TODO for esoteric programming languages
• TODO for greps
• TODO for libraries
• TODO for papers
• TODO for programming languages
• TODO for text editors
• Engines included in and wanted for rebar
• Databases
  • PostgreSQL
  • MongoDB

Research

• Wikipedia Regular expression describes history and POSIX and Perl standards
• Wikipedia Comparison of regular expression engines
• A discussion on Thompson's patent has many leads
  • Glushkov automata [HN]
• A discussion on why intersection and complement are difficult to implement efficiently with large alphabets (Unicode)
• Xerox Research Europe's xfst [HN] [demo]
• Archive and search comp.sources.unix [archive] [mirror]

Exhibits

Extract every engine as a library and wrap each in a consistent C, Rust, and WebAssembly API. Integrate them with Rebar.

Build an interactive demo, that allows you to convert between dialects using InterRegexp and test expressions on text like Regexr, but executing the engines locally via WebAssembly.

Implement each engine with an interactive mode, so you can step through execution of the same pattern with multiple engines at once and visually compare their strategies. This could help demonstrate cases of catastrophic backtracking or showcase engines with efficient algorithmic complexity.

Make nicely-rendered grammars like the regexp grammar in The Racket Reference.

Organize engines into a tree like eylenburg's operating system timeline and family tree, which is generated from CSVs with gnuclad. Solid lines would denote code forks, dotted would denote algorithm or syntax borrowing, and colors would group families (e.g., Spencer, RE2).