#summary Design planning for TTBuffer and friends.

In version 0.2.x, we had the tt_buffer class for storing a bunch of samples and accessing that storage from other objects. When we bring this over to version 0.5.x, what things should we think about?

Here are a few: * What really is the difference between a buffer and a TTAudioSignal? * Assuming that they are really different, how should we handle the implementation of multichannel buffers? * There is a bug in the 0.2.x version where the gaussian function in tap_buffer::fill() does not produce the correct results. * Need to add file reading to tap_buffer. Should we use the LAME library or some other library to ease this work? * Need to finish tap_buffer_record when porting it - in part so that tap.sustain~ can be made into an external (for reliable operation in general, and also at reduced sample rates).