At Tue, 10 Jan 2006 21:56:38 +0900, Nick May wrote:
>
> As far as I can determine, SJIS, in all its hacky, tacky glory, is
> the least cpu intensive of EUCJP, UTF-8 and SJIS encoding schemes.
> Which may be an issue on handsets and other low power devices.
That claim was made in the TRON article, but I don't see how EUC-JP is
any more CPU intensive than SJIS. They both combine ASCII with
variable-length sequences of high-bit characters, and SJIS actually
has more separate ranges to test for the initial byte.
UTF-8 as an encoding is simpler and the processing of each byte should
be faster, however the 2-byte characters of the Japanese encodings
take 3 bytes in UTF-8 so you would expect a performance (and
bandwidth) hit in handsets there.
I don't know anything about the TRON architecture, but the article
itself is overly biased and subjective, and refers only to an ancient
version of Unicode that no longer exists. I'd like to discourage
people from reading it because it provides only disinformation (unless
you already know modern Unicode and are interested in historical
debates).
--
Alex
Received on Wed Jan 11 04:16:03 2006