Re: comments on draft-newman-i18n-comparator-05.txt

This is exactly why systems like Java mandate that NaN be given a 
determinate sorting order.

Whenever you allow partial orderings, almost all commercially available 
sorting algorithms break, sometimes badly, often then resulting in 
corrupted data.

It is only very specialized systems that can deal with partial orders. 
If someone wants to have a separate registry for such types of 
orderings, for special academic purposes, great.

Mark

Philip Guenther wrote:
> Mark Davis <mark.davis@icu-project.org> writes:
> ...
> 
>>Equality MUST be an equivalence relationship (reflexive, symmetric, and 
>>transitive).
>>
>>Ordering MUST establish a total order (that is, < is transitive and 
>>trichotomous), and must be consistent with the Equality relationship.
> 
> 
> These would appear to outlaw a collation for comparing or ordering
> floating-point numbers including NaNs ala IEEE 754 and its successors.
> 
> While collations that lack the above properties can be tricky to use
> correctly, I would be leery of outlawing them given the existence proof
> of widely used systems that lack them.
> 
> 
> Philip Guenther
> 
> 
> 

Received on Friday, 23 September 2005 03:59:58 UTC