home/ atoms/ livecode-keystroke-timestamp-replay

Logging every keystroke with a timestamp lets a live-coded performance be regenerated exactly from text

If a live-coding editor records every keystroke together with the time it was pressed, the entire performance becomes reproducible from that log alone: replaying the keystrokes at their original timings re-runs the exact sequence of edits and therefore reproduces the audio. McLean’s FeedForward has no conventional save function — instead it timestamps keypresses, so a set can be replayed ‘keystroke by keystroke.’ The practical payoff is documentation and recovery: if an audio recording fails, the performance can be regenerated from the keystroke text; and the log is a precise, minimal record of exactly what was played. This treats the performance as a deterministic function of timed text input.

Examples

After a 45-minute set, FeedForward’s keystroke log can be replayed to reproduce the whole performance note-for-note; if the audio capture was lost, the text log regenerates it.

Assessment

Why does timestamped-keystroke logging make a live-coding performance reproducible where a plain final-code snapshot does not? What must be captured besides the characters themselves?

“every key press is actually saved with the timestamp so you you can replay um a performance perfectly”
corpus · alex-mclean-tidalcycles-growing-a-language-for-algorithmic-p · chunk 5