Re: UTF-8S score keeping

From: Tex Texin (texin@progress.com)
Date: Wed Jun 13 2001 - 22:45:41 EDT


Hi Carl,

I agree with you.

My comment on cpu performance was meant to be interpreted in
contrast to the claim
that re-sorting utf-8 data as if it were utf-16 data will cost 1%.
The fact that CPUs increase performance by doubling every 18 months,
or by 1% a week, means that the 1% cost for re-sorting is
negligible.
In the time it takes to have this debate, CPUs have compensated for
this
cost.

However, I should have another bullet under disadvantages indicating
the performance cost for the new counting routines.

I don't really intend to maintain this list however. If someone else
wants to, I am fine with their taking over it.
I'll give up my copyright to it... ;-)

tex

"Carl W. Brown" wrote:
>
> Tex,
>
> >
> > Hardware improves cpu performance 1% in a week.
> >
>
> This is not true. If they are not processing the data they can use UTF-16.
> So why UTF-8? Yjeu plan to use the data in their processing. To do that
> they will have to develop a whole new set of support routines that will run
> slower that UTF-8 because it will take more processing to determine actual
> character length. It is more added overhead than sorting UTF-16 in code
> point sequence.
>
> Carl

-- 
-------------------------------------------------------------------
Tex Texin                      Director, General Product Manager
mailto:Texin@Progress.com      +1-781-280-4271  Fax:+1-781-280-4655
the Progress Company           14 Oak Park, Bedford, MA 01730
-------------------------------------------------------------------



This archive was generated by hypermail 2.1.2 : Fri Jul 06 2001 - 00:17:18 EDT