W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1994

Re: A truly multilingual WWW

From: Gavin Nicol <gtn@ebt.com>
Date: Fri, 30 Dec 1994 10:03:26 -0500
Message-Id: <199412301503.KAA13228@ebt-inc.ebt.com>
To: glenn@stonehand.com
Cc: www@unicode.org, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
>Unless you store bunsetsu boundaries and yomi (phonetic reading) along
>with the original text, you can pretty much forget about automatic
>transliteration of Japanese (or Chinese for that matter).

Well, 100% accurate transliteration, or transcription, like 100%
accurate machine translation, is one of those "hard" problems.

I have seen systems that can take a paragraph of kana and convert it
into kanji, or take a paragraph of kanji, and convert it into kana.
The accuracy is about 98%.

I have also seen machine translation systems working on PC's that can
achieve readable, if not grammatically correct English and Japanese. 

Resource usage for the above systems is still fairly high: about on
par with a typical Microsoft application in fact :-)

Not impossible, not impractical, but not 100% perfect either. Hmm,
sounds familiar...
Received on Friday, 30 December 1994 07:03:04 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:10 EDT