Re: [Gzz] PEG: utf8

[Tuomas Lukka]

On Thu, May 15, 2003 at 01:03:51PM +0300, Antti-Juhani Kaijanaho wrote:
> On 20030515T074958+0300, Tuomas Lukka wrote:
> > > Use Omega.
> > 
> > We got \usepackage[utf]{inputenc} instead. Does omega have benefits
> > over it?
> 
> I didn't know there's such input encoding (BTW, utf is a totally wrong
> name for that, since there are several mutually incompatible UTF's, but
> it's not our problem).
> 
> Googling gives me info on a utf8 inputenc, which is more understandable.
> Is it the same you refer to?  

Yes, I think.

> It is a hack, one which takes a lot of TeX
> memory and supports a very limited set of Unicode characters -
> basically, enough to get the equivalent of latin1, but not much more.
> Omega, on the other hand, is a Unicode-based variant of TeX.

Ok, we'll need to start looking into moving to that then. Shouldn't
be a big problem. Thanks for the tip, I wasn't aware at all of omega's
existence.

> > > > - Are you serious about using smiley faces or other special
> > > >   unicode characters in identifiers?
> > > 
> > > Consider using some appropriate normal form for matching identifiers?
> > 
> > I don't think that's up to us, but up to the language in question..
> 
> Most languages that I am aware use a variation of the Unicode identifier
> format, which AFAIK does not allow smileys in identifiers.  That's why I
> assumed you were talking about some other class of identifiers that
> you've come up with.  

No, just the usual Java stuff. And even if the real smiley characters
are forbidden, there are surely characters in the allowed identifier
characters that will look right ;)

> The recommendation is, BTW, to use NFC for
> case-sensitive identifiers and NFKC for case-insensitive identifiers.

NFC? NFKC?

I'm afraid my knowledge of unicode is still sketchy.

        Tuomas