Sunday, July 6, 2014

Translating typos with Google search

Its probably fairly evident that I love Google's services by now, and I thought I'd just highlight a particularly useful feature in Google search.

When translating documents upwards of 10000 characters (10+ pages long), the chance of finding a typo somewhere is not by any means small. I have on numerous occasions banged my head over a frustrating translation, where a word would simply not make sense in context.

My usual procedure is to double tap Ctrl+C to bring up Golden Dict, into which I've plugged in an offline version of the indispensable WWWJDIC, which will resolve 90% of my queries immediately. If this fails, I alt-tab to my browser where Weblio is waiting with a whole host of dictionaries (including WWWJDIC, and the Life Sciences dictionary by Kyoto University), which will catch the remaining 9.9% or so (and also tends to give some very nice technical usage examples).

If a word fails at this point, frustration sets in as I splice and dice the word to see if the supposed word is actually 2 or more words strung together, but even this will not help if the word is a typo.

In this case, there are few options left available, but a Google search is often invaluable, not only providing some nice usage examples, but sometimes even finding a definition in some obscure internet glossary. In the case of typos however, Google will automatically search using the "correct" spelling. At this point, the fact the search term is actually a typo becomes clear and we can start at the beginning with the correct term*. Lovely! Imagine trying to work this out for yourself with paper dictionaries back before computers were on hand. Eugh.

A word of warning, though. Google love search terms that are common, and it isn't necessarily obvious whether the Google search is deciding against using your search terms because of a typo, or because it is just prioritising what it thinks you want to read about.

*Today I came across "胚葉体形成" in a source document, which can be split into "胚葉" germ layer, "体" body, "形成" formation, which I initially translated as "germ layer formation", but was unhappy with the context, and while dropping "body" improved the flow I wasn't happy about it: "body" could have referred to cells in a germ layer. I decided to Google the term and see if there was a useful precedent, only to be bombarded by results for the similar: "胚様体形成". At this point, I realised "胚様" and  "胚葉" are homophones, and that this may be a typo. Sure enough typing in "はいよう" with Microsoft Japanese IME puts "胚葉" at the top of the candidate suggestion list. "胚様体形成" translated to "embryoid body formation", which fit better into context.

No comments:

Post a Comment