Humanist Discussion Group, Vol. 14, No. 638.
Centre for Computing in the Humanities, King's College London
<http://www.princeton.edu/~mccarty/humanist/>
<http://www.kcl.ac.uk/humanities/cch/humanist/>
Date: Sat, 03 Feb 2001 10:33:06 +0000
From: Willard McCarty <willard.mccarty@kcl.ac.uk>
Subject: black box vs. glass box
This is a question about how we as computing humanists deal with complex
systems.
In physics, as I recall, the term "black box" refers to a device or process
whose internals cannot be known directly. You can measure the inputs and
outputs but cannot see what is inside. As a student of 2nd-year physics at
Berkeley, I was given a metal box with input and output terminals, and with
the aid of a signal generator, voltage meter and oscilloscope was supposed
to discover what was inside. (I don't remember whether I completed the
assignment successfully, but I was certainly intrigued.)
Nowadays programmers checking code have the choice between a black-box and
a glass-box approach. In the former what you do is, as in the example from
physics, use various inputs to generate outputs and by the latter tell
whether the code is correct or not. In the glass-box approach the
programmer checks the code itself, step by step. The latter becomes harder
and harder to do as the code becomes more complex; under certain conditions
it is simply impossible.
Less prejudicially, perhaps, but more analytically the black-box approach
may be defined as
>A strategy for investigating a complex object without knowledge or
>assumptions about its internal make-up, structure or parts. The method
>aims at either a formal description of the transformation rules linking
>inputs and outputs or the construction of a model exhibiting a behavior
>that approximates what is observable from the outside of the "black box".
><http://pespmc1.vub.ac.be/ASC/Black_metho.html>
Now for my point, which is to say, question. Consider the following two
fictitious scenarios:
(1) In the investigation of a complex matter in a large text, a researcher
applies various sophisticated statistical methods to the text. Let us say
that these methods yield very interesting, highly significant results in
the light of what scholars know about the text and from their experience of
it. Nevertheless, the researcher in question does not understand how these
methods work in any detail nor can he or she justify the use of the
particular methods used, rather than others available for the task. All
that he or she knows is that they work, i.e. yield interesting results.
(2) Another researcher, investigating a piece of music, uses various
complex transformation routines to alter the sequence and/or timing of the
notes. Let us say that one of the transformations reveals something highly
significant about the music that no one has ever noticed before. But this
researcher also does not understand the transformation and cannot justify
its use rather than another.
In both cases, how should we regard the researchers' use of these black-box
methods? Are they by principles of good scholarship obliged to determine
how their black boxes work? Could we say that in our domain use of the
black box is fine as a way of getting inspired but that we shouldn't call
the practice humanities computing unless the research goes on to pry into
the box? If we are obliged to understand, then at what level? As in the
definition of primitives, can we say that past a certain point we do not
have to know, but before it we do? Thus, I might argue, since sorting a
list of words is a primitive, I don't need to be able to follow the code
that does the sort. I can, however, imagine someone saying that I should be
able to understand the logic of the process.
Comments? Better questions?
Yours,
WM
-----
Dr Willard McCarty / Senior Lecturer /
Centre for Computing in the Humanities / King's College London /
Strand / London WC2R 2LS / U.K. /
+44 (0)20 7848-2784 / ilex.cc.kcl.ac.uk/wlm/
This archive was generated by hypermail 2b30 : Sat Feb 03 2001 - 05:49:02 EST