Re: Normalization Form KC for Linux

From: Markus Kuhn (
Date: Wed Aug 18 1999 - 16:35:42 EDT

"Addison Phillips" wrote on 1999-08-18 19:53 UTC:
> Old fashioned FACE-like interfaces will not survive KC because they rely on
> half- and full-width forms for positioning.

My idea was not to suggest that a KC normalization be inserted
automatically into the processing pipeline on Linux systems anywhere,
but that KC shall be a guideline, as to which subset of all possible
Unicode encodings should be preferred by new applications. And in this
sense, I think it is not too bad. This does not mean that we will not
support precomposed Hebrew, circled digits and halfwidth forms in xterm,
etc. but it just means that people who wonder about which encoding
alternative to use are encouraged to prefer the KC normalized one in new
software where there are no backwards compatibility requirements. Such a
recommendation limits a bit the number of alternative choices, which
often simplifies thinking about the issues involved for Unicode
newcomers. There are a number of characters in Unicode, which I think
shouldn't be in the character set that we will use in 50 years from now,
and that can only be justified by backwards compatibility, and so far it
seems that normalization form KC removes these quite nicely.

Well, in any case, I have changed the recommendation to Normalization
Form C in the FAQ for the moment.


Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at,  WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:51 EDT