The problem of translation without localization

I recently bought some books from Packt Publishing. I was happy with the smooth purchasing process (especially free shipping to Asia — woot!), and they've made some much appreciated improvements to their ebooks, eliminating the onerous copy-protection scheme that they used to use.

But they really dropped the ball localizing their purchase page:

Prefectures drop-down list

Prefectures drop-down list

The first thing to notice is that the entire page is English, except the drop-down list of prefecture names, which is inexplicably in Japanese. But that's not the problem I want to point out here: the localization problem is that they translated the prefecture names, but failed to put them in the proper order.

The prefecture names appear to be in haphazard order, but it turns out that they're in numerical order, which means that they're ordered by the Unicode values of the Japanese characters. For example, the character 三 in 三重県 has the Unicode value of 0x4E09. Since this is the lowest value in the list, it comes first, and so on.

Of course, this is totally unexpected, because it's nowhere near the dictionary order (the order that it would appear in the dictionary). Numerical sorting works in English, because the ASCII letters are conveniently in alphabetical order, but doesn't work for Japanese, because the same character can have many pronunciations.

But in this case, sorting in dictionary order wouldn't have worked either. Japan has an idiosyncratic but very logical custom of sorting the prefecture names geographically, from Hokkaido in the north to Okinawa in the south. If you have a rough understanding of Japanese geography, this makes finding the desired prefecture easy.

I imagine that an actual Japanese person would find a list of prefectures in this order bewildering, and it would create a great deal of frustration hunting for the proper one.

I think this failure is due to a lack of appreciation of the need for localization. The developers sent off the list of prefectures to be translated, and shoved it into the existing form without due care.

The kicker here, however, is that the list was actually sorted in the proper order, and then "un-sorted" by the developers. Looking at the code for the drop-down, you see that the codes associated with each choice are out of order:

<option value="">Please select</option>
<option value="951">三重県</option>
<option value="953">京都府</option>
<option value="968">佐賀県</option>
<option value="955">兵庫県</option>
<option value="928">北海道</option>
<option value="939">千葉県</option>

If you sort the list by the "value" attribute, it's actually in the proper order, with 北海道 (Hokkaido) first at 928, followed by Aomori at 929, down to Okinawa at 974. So the developers got the list in sorted order, thought, "That can't be right," and proceeded to un-sort it into numerical order…

1 comment to The problem of translation without localization

  • Wow… it’s sad to see that developers overlook such important details… although I didn’t know that the Japanese people ordered the prefectures by “geographical” order. That’s kind of nifty, although a bit confusing and even open for interpretation. Did you tried addressing this issue to the people in charge of that website?

    Thanks for sharing!

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>