Re: An argument against multiple character sets

["Martin J. Duerst" <duerst@w3.org>]

At 12:39 00/01/27 +0100, Patrik F$BgM(Jtstr$B‹N(J wrote:

> It is not so complex as the mapping can be done through some layers of hash
> tables or alike. It is also the case that "we" can descide that the mapping
> should be done on the client side before issuing the query so the server
> only has to do bit-wise comparison.

Agreed.

> So, whether the mapping into some canonical format should be done on the
> client side, or server side before matching occurs, is yet a separate
> question which should be answeered maybe in a form like "more complex
> charsets will imply more complex matching rules and if those imply a great
> increase of CPU cycles, it is better if those are spent on the XXXX side".

I don't think the main point is CPU cycles. One important thing
to consider is that on the server side, you have to be prepared
for every eventuality. On a tiny device, you may not want to
implement this, but you don't have to, you just have to make
sure you don't generate anything non-canonical, which on
a small and controlled device is rather easy.


Regards,   Martin.


#-#-#  Martin J. Du"rst, World Wide Web Consortium
#-#-#  mailto:duerst@w3.org   http://www.w3.org