23 January 2008

The headache that is translating software.

I've never been a professional translator, but I do know it's a thankless task at the best of times.  Take Anthea Bell and Derek Hockridge.  Household names? I think not -- I had to look them up for this article.

Bell and Hockridge translated the Asterix books into English.  This is not a purely technical task, but a creative one.  Asterix relies on puns and wordplays, most of which cannot be translated -- at all -- so it's fair to say that these two wrote a large portion of the jokes that English speakers read in Asterix books.

So, unknown despite their groan-inducing genius, but at least they had coherent material to start off with.

Pity the man who translates the computer program, for to translate something you must first understand it, and computer jargon in English is a right mess.

English is the first language of computing, and most people weren't there when it started.  I wasn't (I wasn't even born then), but I was still a relatively early adopter in the home-computing scene. I had a 32kB Acorn Electron and a 64kB C64.  They low definition screens of 320x 240 pixels, and each letter, number or symbol (including space) was 8 pixels high by 8 pixels wide. Have a look at the specs of your computer -- very, very different.

What I've written so far would have taken up the entire screen of a C64, so when you were writing something (when I was a child I wanted to write games for a living) you chose your words carefully.  With screen-space at a premium, the best word was normally the shortest.  It wasn't just the amateurs that did it -- the professionals were at it too, and the first computers had a lot less than 64k.

Why did they start using the word "error"? Why not "mistake" or "problem"?  The word they chose was predestined to become one of the most common computing terms, and by dropping 2 letters they knew they could save precious bytes, screen inches and printer ribbons (yes, printers had ribbons in those days!)

But what does "error" mean?  The Oxford dictionary lists 6 different senses, several of which are subdivided into multiple further senses.  When you look at your translating
dictionary, are you translating

"3. a. The condition of erring in opinion; the holding of mistaken notions or beliefs; an instance of this, a mistaken notion or belief; false beliefs collectively."

"4. a. Something incorrectly done through ignorance or inadvertence; a mistake, e.g. in calculation, judgement, speech, writing, action, etc."

"4. b. A mistake in the making of a thing; a miscarriage, mishap; a flaw, malformation."

"4. d. Math. The quantity by which a result obtained by observation or by approximate calculation differs from an accurate determination."

Each of these is different, and not all dictionaries make it clear which one's which.

The computer term "error" can mean either a mistake the programmer made in writing the software or a mistake in user input.  Does the same hold true in your chosen language?

Then there's the unnecessarily technical terms that people don't understand.  Did you mean that "delete" means to remove from a list?  No?  Well fine -- you don't need to.  This is just a peculiarity of how disc-drives work and while the first generation of computer users needed to know this, it's not really been very important since the 70s. (Oh yes, and it's a short word.)

There's also the category of analogies and metaphors.  Files, folders, desktops, mice... these are all figurative terms, but they've become fixed, sometimes to the detriment of clarity.  Take files and folders.  I knew someone who would always call a windows folder "a file".  In computer-speak, a file is a single document, and a folder holds multiple files.  However, in the traditional paper-based office where this gent had spent his life, a file is a collection of several documents -- it is a folder and everything in it.  The metaphor is flawed, so a direct translation of the analogy would be equally flawed: ie a waste of time.

Finally, there's the issue of random inconsistencies and redundancies.  Take, for example, the words "sign", "symbol" and "character".  "Character" is a term for the graphical representation of a letter, number or other symbol.  In Windows Vista, € is called the "euro symbol" and the Gaelic translation calls this a "samhla euro", whereas we have a "multiplication sign" -- "soidhne iomadachaidh".  Now there's no difference in Englishy meaning between sign and symbol in this context -- it's simply an arbitrary split of usage.  There is no reason for the translator to maintain a distinction that adds neither clarity not meaning.  I would have opted to merge the terms in Gaelic to a single form. And heck, I'd have included "character" in there as well, as it's just a technological term for any old symbol.
(Incidentally, my mum used to hate us talking about "pound signs" at home. "It's not a sign, it's a symbol," she would say, "A sign is something you see at the side of a road." The same could be said of Gaelic, because "samhlan" is a symbol, and "soidhne" is something you see at the side of the road.)

But that's not the final point, really, because all of the above can melt into one super-sized linguistic problem.

We have the technical term "key", a mechanical device used in musical instruments like pianos and clarinets to make sounds or to modify pitch.  During the industrial revolution, we invented numerous contraptions with key-like mechanisms, but most of these were round, and so the people called them by a word meaning "round thing": "button".

So here we have a technical term and a metaphorical term that essentially mean the same thing, and both of these words have multiple dictionary definitions. A key is the mechanism described above,something you stick in a door to lock or unlock it, a code for secret messages etc. A button is the mechanism, a round thing, or a clothes fastener.

How the hell do you translate that?

I keep meaning to get involved in FOSS translation projects, but every time I look at one I just shudder in dread.

No comments: