[Leica] Test versus trial?, part 1

[imx <imxputs@knoware.nl>]

> * bear in mind that virtually all "tests" in photography are really just
> trials. But that's okay, since all that most of us want to do anyway is try
> things to see how they work for us. Your goal in all of this should simply
> be to learn the way your lenses behave, so that (if you prefer to) you can
> accomodate their weaknesses and play to their strengths when you're
> shooting, and not find yourself unpleasantly surprised by results you didn't
> anticipate. (This is also a goal of learning about other aspects of
> photographic technique.)
The above citation might be expanded and annotated to reflect the true
nature and purpose of several methods of factfinding. Generally all
photographers are experimenters and when they change a film, or a lens, or a
negative developer or a printing paper individually or as a set, they try to
change variables, of which they assume that they will influence the
resulting image in the preferred direction. No one will act in such a way as
to randomly change the variables without any guiding principle.
The goal of any experiment or theoretical study is to explore and explain
the outcome of a photogrtaphic process by finding the true causes that
influence the result in a significant way. Only if you know what effect a
certain change in a varibable has on the result, can we accurately predict
the result and avoid unhappy surprises.
To measure is to know and to know is to predict, it is that simple.
The question then is: if we want to predict and anticipate the results of
changes in the many variables that comprise the photographic technique, how
can we proceed in a reliable way?
If you do not have a clue what causes are influencing the required result,
you will start with a series of trials, changing one variable at a time and
noting the results. To make such a process worthwile, you need to keep notes
and record any change and presumed effect in order to see a pattern emerge.
Without making notes and searching for patterns, you are walking in the
dark.
Given the large amount of photographic literature about every technical
aspect, the trial stage can be skipped in most cases and we can go to true
experimentation. That is: predicting from theory which variables will effect
the outcome and in what amount. This last addition is crucial: knowing that
there is some impact by a variable on a result is not enough: to predict we
need exact information about direction and magnitude of the change in the
variable. We know for sure that a change in temperature of the developer
will affect the  density, by by which amount. The same for agitation rythm:
we know it might have impact on the grain structure, but how much impact and
are there other factors that will negate this effect.
Experiments then are best for fact finding in practical situations but to
make the results worthwhile, we need to do:
a lot of experiments to make sure we have a statistically significant
pattern
make copious notes of all changes during the experiments
change variables one at a time to control and isolate the cause to be
studied

A test is the natural extension to an experiment, with one proviso: you use
instrumentation to measure the phenomena you wish to have data on. And you
use a test protocol to make sure that your measurements are valid.
The upshot is simple: a trial is an insufficient base on which to draw any
conclusion, with the exception of heuristic idea forming.
To give evidential support to these ideas, we need controlled
experimentation and a modicum of theory to guide us.
If we do not proceed according to the time honored rules of fact gathering
and corroboration of theories and hypotheses, we stay ignorant.
Of course, it is every individual's right to accept and support with force
whatever conclusions he is happy to believe in.
If this were the universal approach to photographic technique, however, we
would still be in the infancy of photographic lore, like the ideas that
resolution is the important criterium for optical quality, or that high
dilutions in developers generally increases the definition, or that pushing
increases the speed of a film, etc.
Most of these ideas  were based on evidence so scanty that a nude model
would look fully dressed.
Let us illustrate all of this with the bokeh discussion.
See Part 2      
Erwin