[R6RS] revised Unicode SRFI

Matthew Flatt mflatt at cs.utah.edu
Wed Mar 29 09:18:46 EST 2006


At Wed, 29 Mar 2006 15:47:28 +0200, Michael Sperber wrote:
> - Is the restriction of <= 8 hex digits for scalar values really
>   necessary, i.e. does it hurt not to give a limit? 

I don't know. It seemed like no more than 8 would ever be necessary,
and <= 8 might make life simpler for some parsers. But I would be fine
with removing the restriction.

> - I don't much like the name "char-symbol-part?"---the name confused
>   me as to its purpose.  How about `char-printable-symbol-part?' ?

Ok. Or do we need it at all?

> - In `string-normalize-nfd', there's a double "normal".  Shouldn't it
>   be either `string-nfd' or `string-normalize-d'?

I originally had `string-normalize-d', but I noticed that the Unicode
community usually refers to the format as "NFD" instead of "D". Still,
I didn't change until I went back to

  http://srfi.schemers.org/srfi-75/mail-archive/msg00259.html

where Jorgen suggested `string-normalize-nfd'. I see that it's somewhat
consistent with the Java API, too:

   normalize(s, NFD);

So, I figured that the extra "nf" probably belongs, but I have no
strong opinion on this.


I'll incorporate all your other improvements in the next draft.

Matthew




More information about the R6RS mailing list