A channel is function c which maps from a set of timestamps T to a set of values, V, i.e.\
T can be thought of as the set of times for which the value of the channel c is defined. For the remainder of this work, we assume that:
The set of values V specifies the values that the channel can take. It may be limited to integers, real numbers, an ordered set, or an unordered set.
How does this representation map to reality? A channel intuitively
represents a single attribute's changes in value over time. For
example, in sign language recognition, we may be interested in the
extent of bending of the forefinger. This channel, of course, varies
over time. We may capture information about forefinger bend using some
kind of sensor, which could produce a continuous value between 0
(straight) and 1 (fully bent). Thus, in this case, 
. As time progresses, we could sample the value of the
sensor periodically, say once every tenth of a second, and we might
sample the signal for 25 seconds, or a total of 250 samples. So, in
this case T = [0, 1, .., 249] with a simple mapping from a given
element of T to ``real time''.