Subject: Re: i18n of abiword
From: Paul Rohr (paul@abisource.com)
Date: Fri Jan 14 2000 - 15:11:05 CST
This little manifesto is a general note about the fundamental development
work needed for further i18n of AbiWord, and who's going to do it. Thanks
to all of Pruet's great information on Thai character-handling, I'll use
Thai as an example.
To be clear, though. This description applies to *anyone* who is interested
in adding support for *any* language which doesn't simply use the Latin-1
character set. You know who you are. :-)
the goal
========
AbiWord was originally designed by a few English-speaking GUI fanatics who
know a lot about what it takes to edit and format the kinds of documents we
read and write. Our intent all along has been to do the best job we could
to make our stuff i18n-friendly, which is why we tried to design in stuff
like Unicode support from the beginning.
Our official design goal for this product is that it should Just Work, the
way you expect it to, and we'd like that to eventually be true for every
computer user on the planet.
the reality
===========
However, we do know our limits. Rather than botch the work it takes to
truly support languages like Thai, we've decided to save that work for folks
who really know what they're doing in that area. After all, that's what
being Open Source is all about, right? :-)
Based on the information Pruet has provided, and what I know about the
current state of the code, it looks like there are a total of seven issues
that may need to be addressed to properly support Thai in AbiWord. My goal
here is to identify design issues which may also be leverageable for other
languages. As such, they fall into three categories:
- charset mapping
- combining characters
- line break algorithms
I'll send separate messages with details on each category in a few minutes.
In each case, there are two kinds of work needed -- a general approach which
would work for any language which has that problem, and the details specific
to each language.
For example, if any developers interested in adding, say, Latin-2 support
for Czech, solve the issues in category A in a sufficiently-generic way, you
shouldn't need to do much work in that area for Thai. (And vice-versa, of
course.)
bottom line
===========
I'm a native speaker of English. Unlike most Americans, I studied more
European languages than most (French, some German, and a bit of Russian)
when I was in my teens, but I'm fluent in none of them, and I'm certainly no
linguist. Like most AbiWord developers, I'm far from an expert on i18n
issues. That means that my ability to *solve* i18n problems is drastically
limited. So, I'm not going to try. :-)
Instead, I'll reiterate our standing offer to give whatever help we can to
any i18n-savvy developers who want to understand the existing code so that
they can add these features. We're very, very willing to defer to the
expertise of anyone who knows what they're doing here.
Don't get me wrong. AbiWord *will* eventually Just Work for your language
of choice -- as soon as the developers who know how to solve that problem
provide the code needed!
Thanks,
Paul
This archive was generated by hypermail 2b25 : Fri Jan 14 2000 - 15:05:55 CST