Visual dataflow patching (Max/Pd) builds instruments by wiring boxes, distinct from text-based live coding

Two paradigms sit side by side in real-time audio programming. In visual dataflow patching — Max/MSP (from Miller Puckette’s IRCAM system) and its open-source sibling Pure Data, plus Pd’s modern PlugData interface — you build a working instrument by connecting boxes with virtual cables on a canvas, making signal routing explicit and letting you assemble interactive setups without writing text. In text-based live coding — SuperCollider, TidalCycles, Strudel, Sonic Pi — you type and evaluate code expressions. The interaction models differ: patchers manipulate graph topology directly on the canvas; text coders iterate by typing and evaluating blocks. Historically Max and Pd are important ancestors of live coding but usually placed at its edge rather than at its core, because typed code reads as a legible, projectable performance (the ‘show us your screens’ ethos). Neither paradigm is superior; they suit different mental models and performance styles. PlugData additionally runs as a VST/AU plugin for DAW integration.

Examples

Pd/Max: drag an [osc~ 440] box and wire it to [dac~] to make a tone, re-wiring during a performance. SuperCollider: type { SinOsc.ar(440) }.play — same result, patching versus typing. The French live-coding community describes Pd as ‘programmation visuelle pour l’audio temps réel,’ the main environment for ‘live patching.‘

Assessment

Contrast two concrete workflow differences between a Pd/Max live patcher and a text-based live coder during performance. Explain why Max and Pd are called ancestors of live coding rather than live coding proper, and name one context favouring each paradigm.