From: André Szabolcs Szelp (firstname.lastname@example.org)
Date: Fri Jan 09 2009 - 16:32:09 CST
2009/1/9 Asmus Freytag <email@example.com>:
> On 1/8/2009 6:39 PM, Christopher Fynn wrote:
>> ... there are already cell phones available in Tibet which use a
>> pre-composed Tibetan character set:
> As long as these Tibetan character sets can't actually express something
> that can't also be expressed in the standard Unicode encoding of Tibetan,
> there's no issue here. The requirement is to losslessly convert and
> roundtrip the text, not the code element. In particular, if they are true
> pre-composed characters it should always be possbile to transcode them using
> their decomposition in Unicode and then to compose back on re-conversion.
> The only issue arises, when these conversions aren't unique - as was the
> case with converting from shaped, visual ordered Arabic to Unicode's
> implictly ordered and implicitly shaped Arabic. At that point, pressure
> arose to add compatibility characters for positional presentation forms in
> order to allow lossless transcoding of legacy data.
In practice, a well-defined markup (be it in latin, kanji, katakana or
hiragana) would ensure the (not one-to-one, however lossless)
conversion of information. Like with the Tibetan PUA codes used in
Phones as well...
This archive was generated by hypermail 2.1.5 : Fri Jan 09 2009 - 16:33:36 CST