@low @middle @high vs @minimum @median @maximum

Is there a reason why the percentiles in fluid.bufstats~ are @low, @middle, and @high instead of (the more accurate?) @minimum, @median, and @maximum?

Median, in particular is an obvious one, whereas “middle” (of what?) doesn’t really make sense.

I don’t know enough about stats for the low vs minimum, and high vs maximum, but low/mid/high seem like odd attribute names for statistics.

they are not more accurate, just default values of 0 50 and 100 percentile. You want control of what rank you want, then you need name that are true… the min, max, median would force you to have 0 100 and 50, and you of all people want that control to make it 3 values you want… :wink:

They at least sound statistical. Looking at an attribute called @low gives very little information on what it does. (in coding I typed in @max thinking that’s what it was called, and it wasn’t, which led me to the helpfile and then here).

Is there not a term for this kind of scalable percentiles in statistics?

not that we were aware of when we brainstormed the names… you know, we spend time thinking about these names so usually the obvious proposal has been dismissed for a valid reason! @weefuzzy might have newer ideas but we have to start to change names. lowCentile, midCentile and highCentile is longer and uglier, so I reckon you’d like that? :wink:

You also have a track record for some questionable names, hence the questions/prompts…

I personally like minimum/median/maximum, and at minimum(!) those are the default states, and then you can edit those percentiles. So the min/median/max still makes sense in that context.

I’d take lowcentile/midcentile/highcentile over the current ones though. (then I’d complain about them later, but one step at a time!)

low/middle/high could be referring to anything.

I’ll let @weefuzzy decide - he is the musicianly-interface master after all, and he is able to think of many use-cases.

1 Like

@weefuzzy does have a good sense for names and usability.

(my hunch/gut tells me that low/middle/high was a PA decision)

Collective blame in this instance, so maybe it bears the traces of naming-by-committee :wink:

IIRC, it was I who wasn’t up for <x>Centile because it seemed both like jargon and a finger-full. I dimly remember suggesting floor, middle, ceiling but there being concern that this was too flowery. My problem with minimum / maximum is that they’re not really true except when set to 0 and 1 respectively. Anyway, happy to review this choice with the gang.


You know what they say. The only way to fix a problem created by too many cooks, is to add one more cook!

I do like floor/middle/ceiling as it’s at least more abstract and less likely to be confused with something else (EQ style “highs”).

The min/max, however, I don’t see as being “wrong”, you are just adjusting what those parameters represent (which happen to be min/max by default).

I don’t like floor/middle/ceiling - these words mean other things in very nearby scenarios. The ‘right’ way of terming them would be first, median, third which is even worse though. What’s your aversion to low, middle high?

The terms just seem arbitrary in a statistical context, and in patching around I was looking at my code and didn’t know what I was looking at (for a moment). I was like “high?”.