(non) wave table?

[ccastiglione]

Just had another thought about the wavetable lines.  I may have brought this up
before (a long time ago), but here it is again.  I was thinking that the same
classes which define the global wavetable could be adapted for use by machine
devs who want to store table data not normally thought of as "waveform" data per
se;  but rather as a sample-based buffer for other audio data which is used
frequently.  So that if you wanted, you could pre-calculate (say) a sine
function at machine startup and then, for efficiency's sake, use that table data
from then on to generate audio rather than calculating things out in realtime. 
Granted this is a bad example -- the sine function not being too computationally
expensive -- but, well, it's a thought.  Maybe it would be a useful concept to
hang onto for more complicated audio functions.

I got the idea from looking at some stuff I did in Csound a few years ago, when
I was playing around with it.  It's very common in Csound to buffer function
data in a table and then use it;  in fact it seems to be more common than
generating audio functions in realtime.  Or maybe I'm just misunderstanding how
the whole gen / function table thing worked in Csound.  That was my impression,
anyway.  Anyway, if this makes no sense to anyone then I encourage you all to
forget it and blame the whole thing on hallucinations due to spoilt cream in my
morning coffee.

Here's a very simple alternative version of this question:  is there any reason
why devs couldn't use the same OX_WAV (or whatever) class to create their own,
machine-local wavetable instances if they decided it was advantageous to do so? 
And populate them with any arbitrary data they wanted rather than necessarily
loading in waveform data from a file?