The Community for Technology Leaders


Pages: pp. 5-6

Precise Language

Regarding Neville Holmes's observations in "In Praise of Professional Precision" (The Profession, Apr. 2006, pp. 100, 98–99), for many years I took umbrage at the use of technical terms by people who had only the foggiest idea of what the words meant and often managed to use them wrongly.

  • "Quantum jump" has come to mean a big change, which ignores the tiny energy of a quantum.
  • "High octane" has come to mean extremely powerful with no reference to the control of knocking.
  • "Catalyst" ignores the fact that the catalyst must not be included in the resulting product.
  • "Benchmark" makes no reference to the certification of height above sea level that it always included in my surveys.

And so on.

But I have since become convinced that this misuse of words is part of the improvement and growth of our language, and those of us who know the real meanings of words must suffer in silence.

I note in passing that the French are constantly trying to legislate language precision and consistently fail. I further remark that my father spent his life trying to correct the misuse of "lay" and "lie" to no avail.

Eric Weiss

Neville Holmes replies:

These examples seem to me more like metaphors that have evolved to clichés. We can hope that they will disappear from use, though "benchmark" seems to have evolved beyond its technical use to mean simply "a point of reference."

What I had more in mind under the principle of making distinctions is shown in the misuse of "flexible." Technical people should rather, I would hold, choose between "adaptable" and "versatile," which are of quite distinct meaning.

This is similar to your father's point about the contrast between "lay" and "lie." The faulty use of these words very likely comes about because formal English grammar is no longer taught in school. As a result, young people often don't even know that the words "transitive" and "intransitive" exist.

Regarding the French, I get the feeling that they are trying to legislate language purity rather than precision, a quite different matter.

One of the things that the late Ken Iverson taught me was to use language precisely and never to create a neologism when there is a perfectly good old word (paleologism?) that has the desired denotation.

Among my pet peeves in the computer science world are two creepy terms—I call them that because they continue to creep into papers—for a function's number of arguments. These terms, neither of which appears in The American Heritage Dictionary (AHD), thankfully, are "adicity" and "arity."

I think the former is a bastard-generalization of niladic, monadic, dyadic. The latter apparently arises from a similar corruption of unary, binary, ternary. Neither term makes any sense to normal humans.

Ken introduced use of the term "valence" in APL, from chemistry, where it denotes the "combining capacity of an atom …" (AHD). It is immediately comprehensible by anybody with a high school education, which perhaps explains why computer science types eschew it, as people might otherwise understand what they are talking about.

Robert Bernecky

Neville Holmes responds:

This message provoked me into taking up my Oxford English Dictionary. While I completely agree with the preference for "valence," the OED records "adicity" as early as 1882, defining it as "Combining capacity as an element or non-saturated compound is a monad, dyad, etc." On the other hand, they only found "arity" in Fundamenta Mathematicae in 1968 and define it as "The number of elements by virtue of which something is unary, binary, etc."

In English, "valence" as a term in chemistry came slightly later, in 1884, though it was used in German in 1865, and was recorded to mean a herbal medicine in 1485, and valor or courage in 1604.

I think an aversion to these terms is still well-founded on aesthetic grounds, but I can think of many uglier words and initialisms in common use within computing.

Incidentally, "paleologism" isn't in the OED, at least not in my second edition. Perhaps its use has been blocked by "palaeology," being "the study of antiquities."

Component Markings

Concerning the component markings that Dan Landiss noted ( Computer, May 2006, Letters, p. 6), and which amazed Neville Holmes, let me add this factoid.

I studied EE in the US in the late 1950s. As a hobbyist, I bought surplus components. One batch, apparently with European origins, had strange markings, but I was able to decode them by measuring the components' values. I discovered that these components had the decimal point marked just as Dan Landiss noted. I sent this information to an electronics hobby magazine and they paid me for letting them print that "tip."

Conclusion: These markings were not used in the US, which had its own standards using a color as an exponent multiplier. China and the Pacific Rim countries did not make components in that timeframe. These markings were standard in Europe, as proven by usage.

Observation: Australia was out of touch with the world in the 1950s, not having instant Internet access and cable news. Just because Neville was not aware of this regional standard does not make it a myth. His suggestions remind me of the psychologist that I worked with who invented his own color coding sequence of other items instead of using the one used for electrical components, which is based on the easy-to-remember sequence of colors in the rainbow.

The notation had nothing to do with teletype codes or ASCII. It was more likely a human factor decision to minimize errors or just to simplify the military-driven color coding used in the US.

While I appreciate Neville's elaboration, unfortunately he did not actually clarify anything. He acted as if he knew everything and if he hadn't heard of it then it didn't count.

I found out about the markings by encountering them. And certainly anyone who was not in Europe would have been unlikely to have run into it. But they should acknowledge that fact and qualify their statements, not deny the existence of a major continent and its practices.

My point is that there were no international standards, but there were regional standards. Just because someone doesn't know about other regions doesn't mean they are irrelevant or that their standards didn't exist. Intellectual honesty would demand a simple, IBM-like statement that I don't know, but I will find out instead of promulgating a definitive answer that is wrong.

William Adams

Neville Holmes responds:

First, I confessed myself "completely surprise[d]," not "amazed." Amazement connotes a tendency to disbelief, while surprise connotes unexpectedness, a distinction confirmed for me just now by an inspection of the Webster's Third New International Dictionary in the nearby campus library. I used the word "surprise" precisely because it was something I hadn't known. If using "completely" seemed somewhat journalistic, then I'm sorry for that.

Another thing I don't know is the implied connection between the color codings on resistors and the decimal representations Landiss described. Certainly, back in the 1940s and 1950s, I was familiar with those color codings, as all electronic hobbyists in Australia were at that time, despite our isolation.

Something I do know, something confirmed by my old textbooks, is that Ω was solely, persistently, traditionally, and internationally used as the symbol for an ohm of resistance, with prefixes such as µ used for scaling.

I didn't claim to "know" that ASCII had killed off this tradition; my word was "suspect." In "Toward Decent Text Encoding" ( Computer, Aug. 1998, pp.108–109), I made a simple, IBM-like, statement of my reasons for this suspicion, though I had retired from IBM a decade earlier.

The statement about clarity was in my last single-sentence paragraph, which ended "to clarify and elaborate this point." By these last two words I had meant to confine the claim to the immediately preceding paragraph. I apologize for not having made this clearer.

Finally, I am at a loss to respond to the last paragraph, as it is not clear to me what "a definitive answer" refers to.

64 ms
(Ver 3.x)