The Homonym Trap
One of the most common mistakes made by developers is what I’d call the “homonym trap”. Whether you agree or not with my previous post about the milder form of the linguistic determinism, this is a “built-in” trap in every language, so everybody is suspected to fall in.
What are the homonyms?
Homonyms are words which have the same form (orthographic/phonetic) but unrelated meaning. If they only differ in one way they are called homophones and homographs respectively. In derivation, homonym means “has the same name”; homophone means “has the same sound”; homograph means “written the same”.
From the FreeDictionary.com
Since in coding we don’t really have to deal with the phonetic aspects of the language, let’s concentrate only on the genuine homonyms — having the same name. (Quite mistakenly people usually think about homophones, when they encounter the word homonym.) A real homonym pair in English would be: bear (to carry) and bear (the animal — Ursus arctus or Ursus americanus). Same form, totally unrelated meaning.
Now the trap is that a pair of English homonyms almost for sure would mean two different words in another language:
- bear [verb] - Latin: ferro, ferre, tuli, latum
- bear [noun] - Latin: ursus
As a subgroup for the same mistake I’d mention the words with multiple meanings. Let’s take a very simple example, the preposition to.
to –prep. 1. a in the direction of: Go to the right. b toward and reaching: to fall to the floor.
[...] 19. To is used to show action directed at or toward: Give the book to me. Speak to her.
source: GAGE Canadian Dictionary
Let’s take this last example to the computer world: Send the email to him. No wonder in every email client you will see a line TO with a field to fill in the address of the addressee. For the languages that use a preposition (like English) to “show action directed at or toward”… it’s easy. They just have to find the adecvate preposition in their own language (e.g. An: in German email programs)
Things get a little bit more complicated in the case of the so-called agglutinative languages which use suffixes to express the same meaning. They usually opt to translate this line in email clients with the word addressee — meaning the person to whom a letter, package, etc. is addressed. E.g. in Hungarian: Címzett.
Well, now imagine a script that has a kind of contact form with the TO line (or string) to be displayed, but later the developers also added a small feature that tells the reader “you are seeing the first five articles out of the total 23″. Since this would be a too long phrase, they chose to show something like this:
1 to 5 (of 23)
The only problem is they would use the same (English) string — to — which was originally used in the email form, and when preparing the script for translation they gave only one chance to the translator: translate it in the context of an email address. Which they did. Consequently in the Hungarian translation (localization, if you wish) of the PostNuke you will see a line like this:
1 Addressee 5 (23) — in original: hír 1 Címzett 30 (107 ), where “hír” means news. What the totally nonsensical phrase meant to be: news from 1 to 30 out of total 107. But as it stands now it sounds as if in English you would have: new 1 Addressee 30 (107).
I could bring another example from this very software I am using for blogging. Take the name of the days in a week (abbreviated):
S M T W T F S
You’ll notice that two days have identical initial letters, so the developers in hurry will use only five days:
S M T W T F S
Unfortunately this won’t work in other languages because they need 7 (seven!) different letters to denote the days in a calendar e.g. — so many translators went crazy trying to find the “missing” days.
There is one last subcategory I should mention here. In my mother tongue (Hungarian) the phenomenon is so rare that got a special name: nomenverbum. These nomenverba — as the term suggests — are words that look the very same regardless of their category being verb or noun. In English they are so common that nobody even bothers to name them. Just looking up on my toolbar here in the browser I can see: file, edit, view, help… Are they verbs or nouns? I change the view — it’s a noun. On the other hand: View your site(!) Not only is a verb here in the second case but it is also in imperative mode. Actually the whole problem of the nomenverba is so much more important in the software translation that it might require a separate entry. (And a lesson for translators, too!)














March 11th, 2005 at 9:05 (CST)
Time flies like an arrow.
Fruit flies like a banana.
:)
December 13th, 2006 at 11:25 (CST)
Clear and well written explanation of the concept of of “homonym trap”. Thanks for sharing. Keep it up
Don Lapre is a Superstar
webmaster@j-ams.org
December 13th, 2006 at 13:53 (CST)
[...] Here is the story. On my other blog, that deals with Cross-Cultural issues, De Culturis Mundi there is an article about the homonym trap encountered by software writers, coders. [...]
December 17th, 2006 at 19:56 (CST)
Great story!
But I am not sure about “missing days”
December 17th, 2006 at 20:16 (CST)
The “missing days” syndrome has been in version 1.2 of WP. After that, beginning with the v1.5 it was corrected.
February 7th, 2007 at 8:42 (CST)
Done Lapre a Superstar? I have a full review on my site about the business practices of Don Lapre at http://www.onlyreviews.com/donlapre.html It amazes me how long it finally took for this guy to get in trouble, yet he still won’t go away…incredible.