7

This question is closely tied to two previous questions:

However, I think it is important we reach a consensus on a specific aspect of these (and an answer to the above would probably not get the attention it needs).

Here is my suggestion:

How about officially recommending that people use exclusively: 漢字【かんじ】 or (if they must): 漢字{かんじ}, when providing furigana (or romaji reading) after a word in kanji/kana?

In the aforementioned questions, many different formats where suggested, including the slightly more common 漢字 (かんじ) or 漢字 [かんじ]. Let me explain why we should try to discourage their use and push for a limited number of furigana formats:

  • From previous discussions, it appears that (), [], {}, 【】 are, by a large margin, the main formats used and preferred by JLU users.

  • Most of you probably know about (and use) YOU's brilliant Ruby script. It currently offers inclusive support for all possible ruby conventions. The problem with () and [] is that they often do not indicate furigana... leading to some very annoying false-positives.

  • Further down the line, it is reasonable to imagine SE integrating a very similar solution to the site (as an intermediate step, before even considering a Ruby tag in Markdown). If they do, I don't think we will want auto-conversion of all ()/[] to ruby (which would break many posts).

  • {} and 【】, on the other hand, are fairly unambiguous and have practically no reason to be in a JLU post, other than to indicate furigana. Auto-parsing all {}/【】 in JLU and replacing them by ruby, has very little chance of breaking things.

  • Restricting choices to only two options (albeit purely as an "official recommendation": I am not advocating the flogging and banning of users who choose to use something else) means more simplicity and greater uniformity for new users. {} is pure ASCII, and easy to type for anybody, 【】 looks nicer (and is maybe even less ambiguous), but slightly harder to type for beginners, so having both covers all bases nicely.

  • If we agree on such a convention, it would be rather easy to go back and edit titles/posts to replace previous uses of other formats and make JLU fully compatible with a furigana auto-parser (whether YOU's or a built-in SE one).

Note: Of course, a more complete format such as [漢字] {かんじ} etc. would be even less ambiguous and more parser-friendly (ideal for a Markdown format), however, it has the major drawback of not degrading nicely (i.e. doesn't look good at all without parsing). Plus, it would be a lot harder to get new users to use it immediately... So I vote for simplicity until SE provides support for such a tag at the Markdown level.

So what do you think? Does anybody oppose the idea of adding an official recommendation to the FAQ and generally encouraging this convention through JLU? Any reason not to?

Edits/Updates:

  • Amended suggestion to highlight the preference we give to 【】 over {}

  • To be clear: this suggestion is not about using more furigana, or adding furigana to full sentences (something which would require actual ruby tag support to be done correctly). It is about the best way to format current use of furigana: mostly small words and titles.

  • Regardless of issues of ambiguity/parsability, I hope it will be clear that the central argument in favour of a unified format for furigana is (human) user comfort. Like any other arbitrary convention, the goal of this one is to ensure a bit of consistency in presentation. I am much less attached to which format is picked than the idea of picking one and sticking to it. The current hodgepodge is not serving anyone.

0

5 Answers 5

5

I'd go for dictionary convention:

漢字【かんじ】

For reference, and are usually offered as conversions for [ and ] in most IMEs.

Personally I find 漢字{かんじ} highly unreadable and annoying, but it may be used as a fallback.
漢字 {かんじ} (separated by a space) is recommended in this case.

10
  • 1
    @deceze: I am 100% with you on taste, but still think it more reasonable to offer a simple ASCII alternative (even though 【】 are easy to get with basic knowledge of one's IME, it could still put off complete beginners). How about "Use 【】 or, if you must {}"? ;-)
    – Dave
    Jun 29, 2011 at 8:34
  • If at least a space is used, it may be acceptable. 漢字 {かんじ} Still not liking it though. ≥。≤
    – deceze
    Jun 29, 2011 at 8:41
  • In terms of flexibility, I think both formats should allow one optional space (I'd recommend no-space for and one space for {, but we need to be flexible on support). Pretty sure most people would naturally use a space before { since that's default typography.
    – Dave
    Jun 29, 2011 at 8:58
  • Sounds reasonable. Amended my answer.
    – deceze
    Jun 29, 2011 at 9:57
  • I personally find { } really ugly for this and most of the time I'm not using an IME so I use ( ). I'd say use 【】 if you can and whatever you prefer if you can't and editors should be permitted to change any obvious ascii alternatives to 【】 if they have the urge. Jun 29, 2011 at 10:29
  • @hippietrail: sounds like we all agree on 【】 ;-) But I still think it's important to make a recommendation for the backup as well (and for it not to be ())...
    – Dave
    Jun 29, 2011 at 11:19
  • @Dave: What's wrong with ( )? { } can be very hard to find if you're on a strange keyboard in a strange country. If there's kanji to the left and kana to the right inside ( ) it's not ambiguous. Jun 29, 2011 at 11:21
  • @Hippietrail: () is definitely ambiguous (btw, keep in mind that the above has to work for kanji + romaji and kana + romaji too)... After installing YOU's script, I did encounter at least 2 or 3 instances on recent questions alone (I am a bit too lazy to go back through the site right now, but trust me: they exist! ;-). Problem is: even a few false positives really screw up the experience and justify staying away, imho. Note that all this is nothing but a recommendation, and if anybody for any reason was to use () (for any good reason), I doubt anybody would get angry...
    – Dave
    Jun 29, 2011 at 11:39
  • Yes what is ambiguous and what is ambiguous for a computer script that has to deal with human written text are two different matters. I've had to write such scripts myself. Jun 29, 2011 at 11:41
  • Which is why we want to try encouraging non-ambiguous (for computers and humans) :-)
    – Dave
    Jun 29, 2011 at 13:33
5

Until Japanese.SE supports furigana tags, I think people should use whatever they're comfortable with using/whatever looks best for them. After that, this question will be moot, after all.

I don't think we should be optimizing for people running YOU's script. Remember, no new user is going to be running it!

2
  • as I replied on @Troyen's answer, I still think it worth having an official recommended way, in that 1) it might be a while until SE does something, and when they do, it might be very similar to what YOU's script is doing 2) whenever SE finally adds such a feature, someone will have to go through 100s of previous questions to fix the furigana (unless we take a proactive approach) 3) it's really not all that constraining to use either of these two 4) having a consistent format is an improvement for everybody, regardless of whether they use YOU's script...
    – Dave
    Jun 29, 2011 at 23:57
  • 1
    ... but if for some reason you think 【】 and {} should not be the official pick, definitely go ahead and make your suggestion...
    – Dave
    Jun 29, 2011 at 23:57
4

With the recent embedding of YOU's script, this is an example of what a post can look like:

enter image description here


This is what the post looks like behind the scenes:

example answer with furigana on Safari

10
  • Agreed, this is ugly. (I went back and edited a few posts so we could see this in the wild.) The better solution, as I stated here, is to stop stripping the <ruby>, <rp>, and <rt> tags. This way new users (assuming their browsers support ruby) will see furigana without having to install a script. Jun 29, 2011 at 18:54
  • @Troyen: this is why I started by linking to the two related discussion. Please note that this is first and foremost a discussion on what we can do until SE implements something. Things to keep in mind are 1) it could be a long time (if ever) until they do 2) when they do, it is very possible that they will opt for a middle-ground solution such as what @YOU did, and auto-parse a certain type of bracket into ruby tags 3) when they do, there will already be 100s of questions, it would be nice if we do not have to go through them again one by one.
    – Dave
    Jun 29, 2011 at 23:49
  • Regarding your examples, I think it's obvious that, in the absence of ruby tags, individual furigana in a long sentence is not ideal (whatever format we use) and I wasn't really talking about this. Single words or short phrases (which is the majority of the current furigana uses) was more what I had in mind. In that case, I don't think 【】is particularly ugly: it's the standard dictionary format.
    – Dave
    Jun 29, 2011 at 23:51
  • 1
    @Dave It's not so much an issue of ugliness, but it steals attention. In the example sentences, it's hard to spot that the の is bold because the brackets for the furigana are so much thicker.
    – Troyen
    Jun 29, 2011 at 23:53
  • @Troyen: oh, for the record: personally, I do not use YOU's script most of the time when I am accessing JLU. And I much prefer to see clear furigana markers than the current mixed bag of parentheses. It sounds to me like your position has more to do with "furigana or not?" than "how to mark furigana?"...
    – Dave
    Jun 30, 2011 at 0:01
  • @Troyen: OK... it seems you misinterpreted my original intent. I was in no way advocating a different use of furigana from what is currently done. I also don't think it would be a good thing to mix them inside long sentences (and I don't think anybody is doing this currently). My post only seeks to address current cases where furigana are already used, and how these should be shown.
    – Dave
    Jun 30, 2011 at 0:03
  • @Dave I see. I saw that some people had gone back and added in furigana to all their answers, which I think made them harder to read, and then panicked when I misread the bold line in your title.
    – Troyen
    Jun 30, 2011 at 0:10
  • @Troyen, @Derek: OK, I understand better what you meant. And as I said above, this is not really the point of this question. I am in no way advocating furigana where they don't belong (e.g. long sentences), only better format for existing furigana (words, short phrases, titles etc.). I have edited my question accordingly to make this clear. @Derek: perhaps it'd be good if you reverse your edits, as they do make reading more cumbersome, as both of you noticed.
    – Dave
    Jun 30, 2011 at 0:11
  • @Troyen: well, the use of furigana inside a sentence is a completely separate question. I agree that it makes them less clear. However if they want to do it, I'd still rather they use a consistent format...
    – Dave
    Jun 30, 2011 at 0:13
  • 1
    @Dave: I think we'll have to have furigana in example sentences to be of any use. Students shouldn't have to slog through a mire of kanji they can't read just to get an answer to a question. And even if SE eventually adds ruby support, we won't be able to escape some kind of delimiting pair, whether 【】 or (). @Troyen: I love Chrome, but the font rendering is awful, isn't it? I wonder if we could come up with something cross-browser that also looks good. Jun 30, 2011 at 12:42
2

First, I'm not sure where the best place to put this is, so I'm putting it as an answer to this post. If it is better somewhere else, mods feel free to migrate it.

My concern about this format is how it is parsing the furigana in placing it over the kanji. For now it seems like it's just taking everything in the tags and putting it over the associated kanji. While this works right most of the time, my concern is for beginners using the site. Namely, it would be nice to make a clear distinction between "normal" furigana (where each kanji has an explicit subset of the furigana indicated) and "ateji" furigana, where the furigana is applied as a whole to the accompanying kanji.

For example, 漢字 is a "normal" instance where kan and ji are for each separate letter. Here are two ways to input it: 漢字【かんじ】 and then 漢【かん】字【じ】. My preference (and more correct notation for a beginner) would be the latter because you can clearly see the breakdown of the reading. However, to get it to look this way, I had input it as 漢【kan】字【ji】 (obviously with kana in the brackets). This is clearly a cumbersome way to input furigana, i.e., separate brackets for each "segment" of the reading.

On the other hand, 火傷【やけど】 is a 熟字訓 kanji, where the reading isn't broken down per kanji, but applied blanketly to the whole set of kanji. I wouldn't want a beginner to think it was 火【や】傷【けど】 or 火【やけ】傷【ど】 when this isn't the case.

Now some of you may think this is unnecessary, but I think it could be solved as a simple matter of formatting. I propose putting a delimiter into the "normal" words to denote how they should be parsed out. For example, 漢【かん】字【じ】 could be input as 【かん・じ】, or 日【に】本【ほん】語【ご】 as 【に-ほん-ご】. I would propose that delimiters could be any of {, . / - 、 。 ・ / -}.

This might slightly increase the width of the "normal" words to clearly indicate the correct breakdown (especially for kanji with long readings, like 力【ちから】強【づよ】い), but I think overall it would benefit those using the furigana feature.

Suggestions and comments?

2
  • I understand what you are saying, but I think you mean 熟字訓, not 当て字.
    – user458
    Nov 11, 2011 at 9:52
  • InDesign’s ruby input dialog accepts a space as a delimiter (both zen/hankaku I think) if one wants to enter furigana in group-ruby style (i.e. all furigana for all kanji in a word) but have InDesign display it mono-ruby style (each kanji with its own reading). IMHO a space is easier to type and affects readability less than any other symbol, like in 漢字[かん じ].
    – elmimmo
    Sep 12, 2016 at 8:58
-1

Ok, we've had furigana for a while, and here's my opinion:

*Do not use them*

They are of course welcome when we're dealing with hardcore readings, maybe proper nouns and other fantasies, but I reckon that they should be avoided at all cost for common words. If one wants to know how to read "人", install a damn rikaichan!

Rationale: it's just a pain to read overcharged text full of crappy useless information that cannot even be non-intrusively hidden.

5
  • 4
    Maybe “Do not use them too much” is better. Nov 11, 2011 at 14:56
  • Also discussed a bit in this thread: meta.japanese.stackexchange.com/q/550
    – Troyen
    Nov 12, 2011 at 0:29
  • 1
    Actually, you unwittingly picked a rather bad example: '人' is typically not that easy to read in a sentence for a novice. A dictionary will not always (easily) tell you whether 人 is read じん, にん, ひと etc. Conclusion: I agree with Tsuyoshi: not too much, but still useful sometimes (and doesn't really hurt otherwise).
    – Dave
    Nov 13, 2011 at 4:04
  • @Dave: Well, if you can't read the basic 人, it's unlikely you're going to read anything easy enough for you to understand here. Otherwise, it's covered by the previous sentence (which of course leaves room for using furigana when needed, since it seems to have not been understood properly according to Tsuyoshi's comment and comment's score).
    – Axioplase
    Nov 13, 2011 at 6:39
  • 3
    @Axioplase: I think even a reasonably seasoned Japanese beginner might stumble on something like 大人 etc. My point was just to say that while certain furigana might be a bit superfluous, we'd all have different definitions of what makes them so. Also: it should be a lot easier to hide furigana with 2 lines of custom CSS (doable on any browser I know of) than running Rikaichan, which does not exist for all browsers.
    – Dave
    Nov 13, 2011 at 8:00

Not the answer you're looking for? Browse other questions tagged .