This article turns out to be about modifying software so that it is usable in languages other than English. The authors first take about two pages to say that people who do not speak English will find it easier to use software that deals with their native language, rather than software that works only in English.
They then take a page to say that making such modifications is easier if you have the source code and the cooperation of the software vendor. This is a topic they later return to, explaining that changing software is easier with source code. The alternative is “filters,” pre- and post-processors that translate between the English phrases understood by the programs and the native language of the user.
Next, they make a somewhat useful division of programs into token dependent, communications intensive, and language dependent types. The first describes programs which are driven by commands that resemble English words, since such words would have to be translated. The second describes programs that produce a lot of messages in English. The third describes programs, such as text formatters, that have to know something about linguistic or typographical conventions of the language being processed.
The authors finish up with some examples of trying to modify office software for Spanish speakers. They end up with the unsurprising conclusion that if software is written so that the language-specific parts are separated out, it is easier to adapt.
I suppose that this article might be of some use to people who had never thought about issues involved in modifying office software for international markets. A more useful reference is a pamphlet entitled Software without Borders, distributed by IBM United Kingdom, which better addresses issues such as message length (messages in English always seem to be shorter than those in other languages), character sets, collating sequences, and keyboard layouts.