Re: [icu-support] a question about ICU Segments example

[Manabu O. <man...@or...>]

Hi Yoshito-san,

?

Let’s talk about the example below; this is an example of a label in a chart:

?

当会社のコンセプト

(concept of this company)

?

?

In this example, I like to cut after “当会社の” instead of after “当会社” so that it will show :

当会社の

コンセプト

?

Instead of this: 

当会社

のコンセプト

?

is this possible with the combination of word break and line break?

Isn’t word break not to break “当会社“ into something like “当会“ and “社”?

?

?

Thanks,

Manabu

?

From: Yoshito Umaoka [mailto:yos...@us...] 
Sent: Tuesday, January 26, 2016 8:15 AM
To: ICU support mailing list <icu...@li...>
Cc: Steven R Loomis <srl...@us...>
Subject: Re: [icu-support] a question about ICU Segments example

?


Manabu Okuhara <HYPERLINK "mailto:man...@or..."man...@or...> wrote on 01/20/2016 02:37:09 PM:

> From: Manabu Okuhara <HYPERLINK "mailto:man...@or..."man...@or...>
> To: ICU support mailing list <HYPERLINK "mailto:icu...@li..."icu...@li...>
> Cc: Steven R Loomis/Cupertino/IBM@IBMUS
> Date: 01/20/2016 02:47 PM
> Subject: Re: [icu-support] a question about ICU Segments example
> 
> hi there,
> 
> I have looked at the slide.
> The example string does not have any meaning as a sentence, so I am 
> not able to say any of strict, normal or loose is good.
> I have shared a logic that may work in many cases below in my 
> original email, but there are always exceptions.

---
"It could be as simple as breaking before Kanji and before Katakana as well as breaking after punctuations like “、” and “。”; it may be a good start but may need more thoughts."
---

Is this what you expect?

I guess you're looking at different break type. ICU "line" break for Japanese text is based on JIS 'Kinsoku' rule. It prohibit a line starts with a punctuation character, but it does not prevent a word across two lines.

I guess what you want is a combination of "word" break and "line" break. It looks you don't want to wrap a line in a middle of a word (although I don't think such rule is commonly used for Japanese text). If so, you could use two break iterator instances - one for word break, and another for line break, then wrap a line at a position which is both word break and line break.

-Yoshito