[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: alpha v0.2
At 08:58 00/01/26 +0100, Patrik F$BgM(Jtstr$B‹N(J wrote:
> --On 2000-01-26 12.30 +0900, "Martin J. Duerst" <duerst@w3.org> wrote:
>
> > There is one suggestion I made to the Unicode
> > Technical Committee for a slight simplification
>
> We also did implement it once upon a time, and I don't think it is the TR15
> itself which is difficult. It is more the case that the problem itself is
> extremely difficult thing to handle.
If what Patrick is referring to is the code he once sent me from bunyip,
then that's Normal Form D (decomposed), not Normal Form C (composed).
Normal Form D is a bit easier to implement than Normal Form C if you
fully implement it, but Normal Form C is designed so that most data
currently around is already normalized, so in many cases, you don't
have to do anything.
As for the implementation, in both cases, I agree with Patrick.
The problem itself really defines the complexity of the task at
hand. The actual algorithm is two or three core routines. Setting
up the tables is where you spend most of the time, but no way to
avoid that.
Regards, Martin.
#-#-# Martin J. Du"rst, World Wide Web Consortium
#-#-# mailto:duerst@w3.org http://www.w3.org