Controlled Language 3-99 (September 1999) by Ami Isseroff The Original Article:CL16: Technical Writers Gain Control

Ursula Reuther describes in her article on page 4 of TC-Forum 2-99 (CL16) a need for "...easily readable, understandable, and translatable documentation. She goes on to describe MULTILINT, a tool to automate control of documentation style.

Here, I shall confine myself to a few narrow technical points, and leave the wider implications of this notion for another forum.

1. The stated requirement for documentation that is both easily readable and easily translatable is self-contradictory. Easily readable prose is idiomatic and in some cases colloquial. It takes advantage of "shortcuts" that are intrinsic to each language to reduce word count, such as dropping the definite article in English, or forming a new word by concatenating several words in German. It is written in the syntax, word order and tempo that are characteristic of the particular language. Therefore, it is inherently more difficult to translate than prose composed from a set of rules applicable to a large set of languages.
2. Consider the following sentence taken from the article:

   "So, given the modular design and flexible handling of the system, the MULTILINT approach consists not only of a system which controls the technical writer, but also the technical writer controls individually what is going to be controlled and how this control is realised."

   With respect, while that might be a perfectly good sentence when translated into German, it has certain shortcomings as English: it begins with "So"; it lacks parallelism; it is forty words (an entire paragraph!) long; it uses the root word "control" four times in a single sentence; and it misuses "given" to mean "due to." Those are some 'dry' and formal reasons why it is a bad sentence. The main problem is that it is difficult to read, and that the reader must make several allowances for foreign syntax before deciphering it.

3. The repertoire of rules that the MULTILINT program will follow will be programmed by someone or a small group, and therefore the program will never be better than the style of the people who decide the rules. Consider the effect if only prose such as the example given above is allowed.
4. The article makes much of the ability of the program to allow exceptions. That is, the writer can shut off or ignore rules or warnings that do not make sense. This assumes that the writer will always be alert to catch the problems, and that the problems will always be obvious. The dangers inherent in spell-check programs that automatically substitute "whorehouse" for a misspelling of "warehouse" are too well known to require further discussion. Malapropisms will always occur because software can never understand what you are writing about. For example, when I wrote "In the beginning the Lord God Jehovah created the heavens and the Earth, and he saw...", the grammar check program commented: "They is usually preferred." A writer may not always be alert enough to catch such a problem.
5. The range of correct constructions that the program allows will always be narrower than that permitted by correct grammar. This will result in a tendency to 'standardised prose,' which some people may find desirable. However, the standardised constructions will often be much wordier and less readable than the correct alternatives that are not in the repertoire of the program.

"transline Deutschland - Übersetzungsdienst für technische Übersetzung"