Take a look at the Japanese text in the picture – this is from a Seoul subway card recharging machine I saw when I was in Korea earlier this year, with the language set to Japanese. Does anything look wrong?
If you’re not a Japanese speaker, you probably don’t see anything off, as there aren’t any blatantly corrupt characters or boxes. However, there is an issue in how the text is displayed that I tend to see from time to time: the Chinese versions of some of the characters are used, which differ slightly from their Japanese counterparts. In the picture this is the case for 直 (and 入 to a lesser extent), which are going to look off to a Japanese speaker (for context, they appear on the button that translates to “direct input” – where pressing it lets you type in the exact amount of won you want to recharge your card with).
Why does this happen?
There are quite a few characters where both the Chinese and Japanese versions share the same Unicode point. This means the version that actually gets displayed is dependent on which font is used to render the text, and if the developer isn’t careful, the software will default to using a Chinese font for the Japanese text.
You can see this for yourself directly if you copy the characters into Google Translate – switching the language between Chinese and Japanese will reveal the change:
Why does this matter?
Because it breaks the seamless experience that users deserve to have when using a product, and in a lot of cases this could turn users off from making a purchase on something they otherwise would have bought. One of the sites I found about the topic gave a great analogy:
iҭ’s liκє if you wєrє rєading somєҭhing in єnglish and iҭ looκєd liκє ҭhis.
Sure, the meaning gets through just fine, but you know something’s weird and it raises suspicion. If you can’t trust the creator to display your language’s text correctly, would you trust them with your money?
Okay, so how can this be avoided?
In many situations, you can specify that a page’s language is Japanese and list out some fonts the system can fall back on, which will ensure the correct versions of the characters always get displayed. More importantly, though, this issue highlights the importance of a step in the localization process called linguistic quality assurance (LQA), where a native speaker views the localized text in-context to make sure nothing is wrong. Many people don’t realize that text translation/editing often occurs in a separate tool where you don’t get to see how the text actually looks like in-context – it’s only after the translated text is compiled back into its original format that issues like this will become clear, making the LQA step a must.
Learning more about this industry the past couple years has made me appreciate localization so much more, and when I see things like this “in the wild”, it serves as a reminder that our work is always going to be relevant. We always need to be on the lookout for opportunities to emphasize the importance of localization to others so we can prevent these kinds of issues from happening!
