Making your own translation tools

I am a profoundly lazy man, in some respects anyway.  I hate pointless labour.  And what can be more pointless than the way many of us translate?

Imagine getting a French text in front of you.  The process goes something like this:

You read the first sentence.  You type an English version into Word.  Then you look back to the book.  A few moments of searching along the line, and you find the second sentence.  You know most of the words, but not all, so you type in a couple of them in an electronic dictionary.  Then you look back again at the page, to get the whole sentence, and spend time again fumbling for it in the mass of text.  Then you write another sentence.  And so on.

Frankly all this switching to and fro is annoying and pointless.

What we need, surely, is to turn the French into an electronic form, split it into sentences, and put each sentence on a separate line.

We could go further.  Machine translators for French are quite good.  Let’s run the electronic text through one of those.  Then split the translation into sentences, and interleave them with the French.

Won’t that be much easier?  We no longer have to find a text in a page in a book; it’s immediately above the line.  We have the machine translator’s vocabulary; that will reduce the amount of looking up.  In short, it’s easier and quicker and less painful.

I’ve written a little utility that does the splitting into sentences and the interleaving.  I use it with a machine translator, and just paste the output back into my utility.

Of course it’s limited in what it does, but the output is a  nice word document with interleaved French and English.

It’s making working on Agapius much easier anyway!  If only there were some way to hover a mouse over a French word and get a full dictionary entry.  Are there any French dictionaries in XML form?

Share

2 thoughts on “Making your own translation tools

  1. This strikes me as similar to the idea of a bitext, and there are some other utilities (e.g. bitext2tmx) which can take the source + automatic translation, align them, and additionally output them in a translation memory format which you can then in turn use with some (possibly free) assisted translation utilities.

    As for French dictionaries, Wiktionary may be worth a look (either http://download.wikimedia.org/frwiktionary/latest/frwiktionary-latest-pages-articles.xml.bz2 for French-French, or en with some special processing to cut it down to only French words and get a possibly lackluster French-English). There’s also FreeDict but I get the sense that for French it is probably not nearly as developed.

  2. Interesting and thank you! I will look at these links. We really do need lots of free dictionaries online, I think. The possibilities for mashups from them are considerable.

Leave a Reply