30 May 2010

Unpleasant truths

Analog slide rule versus digital calculator



The slide rule was the most important calculating instrument from 1715 to 1975. It has been claimed that no significant invention or scientific ida has been developed without the use of the slide rule. The beauty of this instrument is the fact that it is fast, quite accurate and helps to understand the inherent relationships and fundamental rules of numbers and mathematics in a truly visual way.
When HP introduced the first digital calculator in 1975 at a price worth a six month salary of a median worker, the end of the analog slide rule happened overnight. Within two years the slide rule had disappeared and is now a collectors item. The proponents of the digital calculator hoped that the easy access and simple method of calculations could foster the grasp of mathematics and numbers. But the opposite happened! Most students do not have an idea what they are doing when mechanically doing their button pushing to get the required answers.

The analogy with AgX photography and digital image making is evident. When the digital camera was introduced it cost a fortune, but within a decade AgX technology was obsolete and history repeats itself: hardly anybody nowadays grasps the basics of photography or the inherent relationships between scene and image capture and print. You win, you lose might be the comment, but there is more than just embracing a new technology. A great cultural loss may be the consequence if we forget the rules of AgX photography and its role for the method of digital image making.

Tolerances


The cult of numerical testing and comparing numbers is losing its potential for enlightenment. A cult is always in danger of creating excesses. Since the days that Zeiss took the step to publish MTF graphs as a supplement to informed lens appraisal, there is now a flood of numbers with a high level of numerological precision. A lens/camera combination has a resolution of 1221 lp/mm and this compares unfavorably to another lens/camera combo with 1240 lp/mm. The conclusion is quickly done. The internet has only 15 minute of fame for any product and comment. We do forget that any camera and lens has at least a production tolerance of 5%. We also forget that the margins for strictly controlled lab testing is at least 5% too. We also forget that human errors might add another 5%. If we for the sake of the argument stick to the minimal 10%, we see that the first result has a margin of error of 1209 to 1231 and the second one lays between 1228 and 1252. At most! In fact both camera/lens combinations might perform identically.
Now add another variable: the photographer. This variable may account for a margin of -25% to +10%. This number may be surprising. Given the number of parameters (expertise, experience, accuracy, exposure, handholding, focus errors, post processing, in camera processing and so on) this estimate may be even too low.
If we add these tolerances, it becomes clear that 90% of the discussions in the internat forums are simply discussing quality differences that are in fact within the margins of error. That is why these discussions never end and why they are not informative. If one comment claims that a camera/lens sucks and another claims that this same combo is the best there is, both claims might be right within the tolerance range and personal evaluation. But both claims are not informative as both do not take into account the margins of error. The test reports of let us say PopPhoto and DPReview and Color Photographie would become more interesting if they would add their margins of error.

One liners versus true discourse


In the Netherlands we are involved in a vote for a new parliament. Politicians are fond of one-liners that are intellectually empty and polished statements that evade addressing the really important questions and themes. Spin doctors are more important than informed persons who dare to refer to the true problems of our time. And the victim is the classical discourse: a dispassionate analysis of the real facts and what to make of them, the willingness to assume that one could be wrong in the discussion and the idea that a solution is more important than one’s conviction to be right.
This political culture is alas now the dominant force in internet discussions and websites. The internet once was invented to facilitate the free and fast exchange of scientific results and topics where scientists could present results and ask for informed comment.
Current internet mores are far removed form this ideal. And it has missed a great chance to really change the world.
Am I a neo-romanticist?