Messages in this thread | | | Date | Wed, 06 Oct 1999 03:17:18 +0000 | From | Steve Underwood <> | Subject | Re: Unicode support on VGA console. |
| |
Alex Belits wrote:
> On Wed, 6 Oct 1999, Steve Underwood wrote: > > > language (about the same number can speak English and Mandarin, but less > > Mandarin speakers can read), but still much of the human race would like to > > manage low end products in a language other than English. > > > > Promote multi-lingual everyone! > > > > Most people (more precisely, almost everyone) use their local > standards, not Unicode. Unicode is promoted by English-centric or > Europe-centric (iso8859-1) people, and is hated by almost everyone else.
I think this is almost, but not quite, spot on. Until recently I was one of those haters (I am a native English reader, but use Chinese a lot). Only an ASCII-centric user would think UTF8 is anything but a total bodge. Its about as clean and elegant as the Windows 98 code base. It makes all our Asian language text files 50% larger, and until version 3 had some serious limitations. Unicode has been almost universally ignored, as a waste of time, in Asia. It solved nothing and created new problems. Behind the scenes Unicode it is actually used heavily in Asia, as most MS code works in a messy mish-mash of Unicode and local codes (scratch beneath the surface, and you will see the idea of NT as a clean Unicode environment is a bad joke). I recently had a change of heart on Unicode, for a couple of reasons.
However much people hate Unicode (with excellent reasons) the limitations in other codes are no longer being fixed. It is Unicode which is now making progress. For example, no proper standard for adding Cantonese Dialect characters to any other Chinese character set exists (apart from a Hong Kong Government spec. that is not widely used). There is now an extensive set of Cantonese characters in Unicode 3. The nastiest of the merged character issues seem to have been cleaned up in Unicode 3. The character set of Unicode 3 actually seems superior to most others, so the baseline of it being usable has been reached for me. I think its also pretty usable for most other languages these days. I think its inevitable that UTF8 is going to be the form in which most Unicode text exists, and non-English users are going to have to face the text bloat it causes.
The other main issue is of common ground. I think those of us who want to see effective internationisation/multinationalisation want to see it happen at source. I mean we want developers to develop in i18n capability, and have all the translations, etc. in the main source tree. That can only realistically happen with a common character code. Developers cannot understand every language, but they need to be able to display every language, and check their software behaves in a sensible way (90% of this requires very little knowledge of the language being displayed). Right now, this hardly ever happens. Large numbers of people are beavering away converting endless programs to Russian, Chinese, Japanese, etc. At the next software release the work will start all over again. The GNU tools provide a framework for dealing with the bare text translations, but little else that is language dependent. The only realistic way to assist authors to handle the bulk of i18n issues themselves is to use a common character code. The authors can then accept submissions for translation, character input and character display issues, etc. and merge them with the main source tree.
So, I have come to see Unicode as the only realistic way to reduce the reinvention of the wheel in i18n/l10n/m17n issues, and promote the production of properly internationalised open source code. Any opinions to the contrary are welcome.
Steve
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |