[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Conference at debconf on i18n/l10n: ideas? suggestions?


From: Javier Fernández-Sanguino Peña <jfs@computer.org>
Subject: Re: Conference at debconf on i18n/l10n: ideas? suggestions?
Date: Sat, 12 Jul 2003 13:36:42 +0200

> I will add these examples.

There are several unknown problems on i18n around Debian.

1. Package description pages with DDTP

The first character of the Description: is displayed in large font,
but this algorithm is not multibyte-compliant (assumes one character
is one byte).  It results in broken page like:

It is already reported and Frank Lichtenheld is now working.

2. Debian webpage search page cannot search east Asian

For example, the following URL searches Japanese word "news" which
should hit many pages.
The reason of this problem is somewhat special.  Though mnogosearch
has potential to support east Asian languages, the default compilation
(both upstream level and Debian package) disables.  All what is needed
is to compile mnogosearch with one additional configuration option.
This should be fixed immediately but so far I have not succeeded to
persuade Craig Small (maintainer of debian search page).

Another cause of this problem is the assumption of "words are separated
by whitespaces".  I think you will easily understand that search engine
for non-whitespace languages are much difficult than whitespace languages.
Examples of such non-whitespace languages are: Chinese, Japanese, and
Thai, as far as I know.  (Chinese and Japanese are languages where
translation of Debian webpages are available.)

mnogosearch supports chasen, a software to analyze Japanese text and
extract words.  However, there seems a license problem on chasen.

> Could you also please translate
> the debconf mojibake page [1].
> [1] http://www.debian.or.jp/~kubota/mojibake/debconf.html

The page is in English, isn't it?  One possibility is that the
content-negotiation isn't working well.  Please try

Tomohiro KUBOTA <kubota@debian.org>

Reply to: