I think the complexity of determining sample length, before it’s actually being played, is best illustrated with an example…something like mat-weasels scratching demo (link to thread + download)
Imagine that the scratching samples was streamed audio (audiotrack) instead of short samples - and that we had some sort of visual indication of the length? That is, before we actually start to play? That’s where the pre-calculation would have to be done. And to display the length as a simple “bar”, as opposed to a waveform, will not solve anything, as the pre-calculation would still be needed (from a programmers perspective, the waveform would be the last, and relatively simple step of updating the display).
For all of these reason I think we’re better off using a length/waveform display for frozen tracks. Here it makes a lot more sense.
Heck, those waveforms even look like icicles to me (== frozen audio:-)
But it doesn’t display the “mangling”, just the sample!