On 02/09/2001 05:41:22 AM "J M Sykes" wrote:
>I was, rightly or wrongly, regarding as a basic
>principle: that the NF of the result should always be that of one of the
>operands...
I can understand going in with that premise. In the table I proposed, there
is one case in which this doesn't hold: (NFKC, NFD) -> NFC.
>Of course, this could hardly be said to be a matter of logic, more
>the sort of thing users might expect.
Well, I'd wonder how much most users would even be aware of the issues.
But, the principle isn't unmotivated; I think it just wasn't applied
carefully enough. The normalisation form is probably best thought of as a
*pair* of "orthogonal* properties. With this view, we can see that for each
property the result does have the property value of one of the input
operands: If either input is C, the result is C; if either input is ~K, the
result is ~K.
>Thanks again,
You're welcome.
- Peter
---------------------------------------------------------------------------
Peter Constable
Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
E-mail: <peter_constable@sil.org>
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:18 EDT