Re: [xsl] Getting the Base Character of Character with Diacritic

Hi Jeff,

This, I think, is far from trivial, but here's a possible approach you 
might want to consider.
First of all, a character with diacritic does not necessarily need to be 
encoded as one character. For instance, the character Ẫ can be encoded 
as x1EAA, as x0041x02C2x0303 (A + ^ + ~), as x00C3x0303 (Â + ~) or as 
x00C2x02C2 (Ã + ^), where the ~ and ^ are not the characters on your 
keyboard, but the combining diacritical marks. Also, depending on how 
you look at it, Æ (x00C6) equals (is combatible with) AE.
To normalize this Unicode stuff, the Unicode Consortium has invented 
four (or five?) normalization algorithms. These infamous algorithms 
either try to decompose (D) as much as possible, or compose (C) as much 
as possible. Furthermore, they can try to (de)compose even further to 
the compatible (K) counterparts of the characters. This gives four 
variants of Normalization Forms: NFC, NFKC, NFD, NFKD (see http://www
.unicode.org/unicode/reports/tr15). The XPath 2.0 function 
normalize-unicode is used for dealing with this and also adds a fifth 
variant: fully-normalized, which is NFC plus not starting with a 
combining character.
That's for the theory. Now practice. Processors only need to support one 
normalization form, namely NFC. This is needed so you can correctly 
compare two strings. What you need, is NFKD (or, to a lesser extend, 
NFD). Then you can translate these diacritical marks to something 
(x02C2, x0300, x0301m x0303, x0309, x0329, x2C9 for a start: circumflex, 
grave, aigu, tilde, hook, dot, macron) and you have a string without 
diacritical marks.
As far as I can tell, your best bet is Saxon-SA, it seems to support 
NFC, NFD, NFKC, and NFKD 
(http://www.saxonica.com/conformance/xqts100/SaxonResults.html). Not 
sure if it will work the way you expect it to.
Not sure how you tend to deal with Chinese, Hangul, Hebrew, Cyrillic and 
some other more complex languages, because the combining characters are 
an important part of the letter, without which they mean nothing. This 
depends on what you are actually after.
Cheers,

Abel Braaksma
http://abelleba.metacarpus.com


Jeff Sese wrote:
> Hi,
>
> Is there a way in xslt for me to get the base character of a character 
> with diacritic?
> Like ā to a? I was thinking of using the translate function, but it 
> there are too many characters to include.
> -- Jeff

--~------------------------------------------------------------------
XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--