L2/09-233

Source: Mark Davis

Date: July 7, 2009

Subject:  Normalization Properties

 

There were two open issues in the handling of some of the new normalization properties.

1. Whether to move NFKC_CF and/or INFKC_CF into a separate file. If we did that, it would make sense to also include FC_NFKC_Closure, since that is also a Case+Norm property. At this point, I'd rather keep them there; the two Derived.*Prop.* files are about the same size.

(If we were to move properties around, I think what would make sense would be a cleaner restructuring, like:

DerivedCaseProps
DerivedNormalizationProps
DerivedNormCaseProps
DerivedCoreProps // everything else

But I'd rather not do that without UTC buyin)

2. The statement under NFKC_CF

> #  When this property is used to transform strings, NFC *must* be applied
> afterwards.

Ken wants to remove this. I think we really need it here, and in #44; this is an important health warning. It doesn't hurt to include it, and may prevent people from stupid errors.

Ken's response was:
I want to remove it from where it is now, anachronistically
following the specification of the derivation of the
property.

It could reasonably be moved up above the derivation, to the
description of the property. But if it is going to stay in
the comment section in the data file, I would like a better
description of the property. Having a normative claim about
the use of a property like that, without better context for
it, strikes me as a mistake.

We agreed to push to UTC discussion.


Mark