I did open up the other thread. SCDoc: Events/patterns, and searchability of non-class docs
But I also think I should dispel this one right away:
Pbind is a way to populate values into an event. There is intentionally no restriction on the event keys at this stage. I can populate 10 values under non-reserved keys, and use these to calculate a degree or midinote or frequency. This is an absolutely critical feature that we must not remove, under any circumstances.
When the resulting event is performed, there are two senses in which keys are relevant. One is for the built-in calculations (degree, note, octave, scale, etc. have specific meanings in this schema). The other is SynthDef arguments. Keys are accessed from the event for both of these. Keys may exist in the event that are relevant neither for calculations nor for synth arguments. These are just ignored.
If keys are arguments in an event-type class, then it muddles the distinction between data preparation and data performance, by limiting the data that one is allowed to prepare to permit only those data that will be directly performed. This would fundamentally change the nature of the pattern system.
So, no. I would definitely not be in favor of this.
hjh