Re: Compatibility requirements

[Paul Hoffman / IMC <phoffman@imc.org>]

At 02:31 PM 1/23/00 +0800, James Seng wrote:
>Canonicalization algo is an ongoing process.

Not for all character sets. It is fixed and done for Unicode, for example. 
See Unicode Technical Report 15.

>  Upgrades and changes to the algo
>is unavoidable altho we can define some basic rules. Therefore, there will be
>times when the algo is different on different client, say v1.0 and v3.0.

Then those character sets can't be used unless we say that for any 
particular character set, canonicalization must be done at all resolvers 
and master zones (and caches...) using the exact same algorithm and that 
when the algorithm changes everyone updates simultaneously.

You can see the scalability and stability problem here.

--Paul Hoffman, Director
--Internet Mail Consortium