Re: charset question (Greek)

> On Wed, Mar 13, 2002 at 03:14:33PM +0100, Andreas Fink wrote:

> > well in theory this should be the case and it probably is for non
> > european languages but for historical reasons there are some
> > duplications.

Duplication is in the eye of the beholder. Or rather his language.

> > examples:     		ä (German umlaut "a")
> >
> > iso8859-1:	E4
> > unicode:	00E4   or 0061 + 0308 (a + ¨)
>
> Unicode defines normalizations for this.  You can convert Unicode strings
> to "canonical forms" for comparisons.  There are several levels of canonical
> forms, for example there's one that ignores case distinctions.  Unfortunately
> doing this correctly requires the use of large tables, so we might be better
> off linking to a unicode-support library.  There are a few out there.
> libunicode, for example :-)

Or CCC. But beware of the fact that this normalization is generally not
trivial, nor reversible - and for approximations often needs either source
or target -language-. Example: latin-1 0xFF: the y\: which some dutch
people would rather see is a ligature like ij.

Dw