I am using a list of keywords to put into a meta tag for a localized Japanese website (based on an English one). The site is for the Japanese regional branch of a client, so the native speakers from the branch have translated the keywords themselves for us. When we sent the list to them, we had it organized with commas, as one would typically do for keywords:
foo, bar, baz
However, it seems that the Japanese language (in which I have pretty much 0 expertise) has its own comma character, and they used that character when translating the list of keywords. So something akin to the above (from Google translate, used purely for example and not for translation accuracy, for curious people I used fu instead of foo) would come out as something like:
フー、バー、バズ
This uses the Japanese comma character, 、, instead of a normal Latin comma, ,.
Will this affect how the keywords are used? Is 、 preferred for separating the keyword tokens for Japanese-targeted SEO, or is ,?
I searched through Google for some hint at what to do, but any pages I found dealing with localized keyword text were either using a Latin-based alphabet (like French), or for a couple of Japanese ones I found did not actually display any examples that might have even suggested which comma character to use (they really only talked about not using literal translations, which we've already done by having native speakers translate the content). The one place I found with an essentially duplicate question to mine was the forum question posted here, but it has no answers (and isn't likely to get any since it's 1.5 years old...).
Note: I've seen talk about the lack of use of keywords by SEO engines. The client wants keywords, though, so we will be doing keywords, meaning there's little use in bringing up this point in comments/answers.
The keywords have to be separated by the , character, no matter which language the keywords are in. For keywords it is defined that the value "must be a set of comma-separated tokens", which is defined as:
[…] a string containing zero or more tokens each separated from the next by a single "," (U+002C) character […]
Note that this , is not part of the keywords. It's like a reserved character. If a keyword itself should contain a ,, it would have to be encoded (for example as ,).
If you hand over keywords for translation, you shouldn't include the separator character (unless it is part of the keyword itself).
So better send the translator a list like …
- foo
- bar
- baz
… instead of "foo, bar, baz".