[
Lists Home |
Date Index |
Thread Index
]
Roger L. Costello wrote:
>
> That leads me to another question - is feedback generally applied to the
> "data" or to the "process"?
>
Feedback is part of the processing of the data. Think of an electronic
amplifier. The unmodified input signal is the data, the amplifier
provides a transformation of that signal to create the output. Thus,
the stylesheet would be analogous to the amplifier (or other electronic
circuit).
With feedback, a part of the input signal is taken from the output. If
the feedback adds linearly to the input, and the feedback portion of the
output is linear, the system is a lot easier to analyze than if there is
non-linearity.
Thinking about an xslt transformation, its functional design should
prevent feeding back part of the output to the input. Once the input
has been read, it cannot be changed so far as the particular
transformation is concerned. Thus, xslt operates somewhat analgously
(is that a word??) to an open loop amplifier.
So it would seem that to simulate fedback behavior using xslt, you would
need some kind of supervisor sitting outside the stylesheet, capable of
routing both its output and the original input to another
transformation. Could that supervisor also be a stylesheet? I doubt it
but maybe someone else could be clever enough.
Cheers,
Tom P
|
