What is the correct approach to romaji, kana, and/or kanji in example answers?

Question

I like to offer example sentences when I answer a question. However, I can't assume that every reader will know every word I want to use.

How should I indicate kanji readings?

For example, see my response on this question, where I haven't given any readings on 改札口 or 階段, assuming that the reader may be using Rikaichan or Rikai Browser.

For certain questions, the level of the question being asked should give an obvious context for the level of the reader, but not always.

Shall we discuss a standardized approach?

Answer 1

I'm all for automated furigana (with the option to edit the furigana if automated one is wrong, they are 98% of the time correct though). However I would discourage use of romaji simply because it does not help learners at all and we should have a policy of converting romaji into kanji/furigana if possible.

Answer 2

Linguists, including native speakers of Japanese, tend to use romanization when discussing Japanese morphology. Why? Because it makes analysis simpler. Compare the following:

1. The verb 折る has five surface forms, 折ら・折り・折る・折れ・折ろ.
2. The verb or- has one form, or-.

The romanized version looks simpler to me!

What if I want to describe the contraction in 言えば → 言や generally? I can say it two ways:

1. A kana in the え row, when followed by ば, becomes the corresponding kana in the い row, in which case ば becomes little ゃ, unless that kana is え, in which case え disappears and ば becomes big や.
2. -eba becomes -ya.

The romanized version looks simpler here, too!

Sure, Japanese is usually written in kana and kanji. But there's a point to using romaji some of the time, so we should use it when there's an advantage to doing so.

Answer 3

I am all for an "auto-parse" option on the server side (please, do not assume everybody has Rikaichan/kun running: awesome as it is, it only supports a small subset of browsers and I don't think it is the place of SE to tell the user what browser they should be using). It is doable but frankly unlikely at this stage...

In the meantime, perhaps the FAQ could be amended with a short "best practice" recommendation on providing kanji reading. Of course we would first need to come to a consensus, which might not be easy ;-) Personally, I like the form: 漢字【かんじ】 (death to romaji!) for single word mentions, but I have seen many people use 漢字 {かんじ} in questions, and it looks OK too (perhaps easier to type for beginners?).

For what it's worth, HTML already has a specific tag to display furigana (or equivalent) in Asian languages:

<ruby>漢<rt>かん</rt></ruby><ruby>字<rt>じ</rt></ruby>

which is meant to look something like:

かん じ
 漢  字

(with better spacing, of course)

It will display nicely in most modern browsers (including IE, I think). With a bit of extra CSS, it will also display fine on less modern ones.

Unfortunately, SE currently doesn't support the <ruby> tag and strips it from submissions. But I suspect it will be much easier to request they allow it (and perhaps even provide a tool or code shortcut to make its writing less cumbersome) than to have them auto-generate furigana.

Answer 4

The first thing that I can think of is to have some sort of furigana reading but that would likely require a high degree of customization for the site, even if limited to a mouse over that could be supplied. We might be able to get it added to the site at some point in the future if the beta proves to be highly successful.

That said though, maybe having the reading in parenthesis or just supplying a link to the word in an online dictionary (i.e. 読み仮名)?

Answer 5

I prefer this form for example sentences:

ローマ字どうやって書けるかな？ Rōmaji dōyatte kakeru kana?

Inside a sentence, the first use of a 漢字 kanji should be annotated by romaji, any subsequent use of the 漢字 is fine by itself.

Maybe this is too komakai though, I don't think we can get everybody to standardize on the same form.

Answer 6

I think that assuming that the user is using rikaichan is fair.