unaccenting letters

[David Alex Lamb]

Has anyone developed a method to un-accent accented characters? I mean
for example translating e-acute, e-grave, and e-circumflex to e?

 

[Guest]

Has anyone developed a method to un-accent accented characters? I mean
for example translating e-acute, e-grave, and e-circumflex to e?

If using this for comparisons, use a Collator.

I haven't found code for what you ask for, but I assume the algorithm
would be fairly straight forward.

First, do a full decomposition on the character in question.
Second, grab the first character from the resultant String.

You will need reference to the Unicode Standard to create this
decomposition table.

HTH,
La'ie Techie

 

[Thomas Schodt]

Has anyone developed a method to un-accent accented characters? I mean
for example translating e-acute, e-grave, and e-circumflex to e?

I posted something that implements this for the upper half of Latin-1
on the Sun Forums; the last post on this link:
http://forum.java.sun.com/thread.jsp?forum=31&thread=483768&range=24

 

[John O'Conner]

You will need reference to the Unicode Standard to create this
decomposition table.

What you want is already in the JDK...just not documented and not in the
public API. You should check out sun.text.Normalizer. You can create
"normalized" decomposed text. See the Unicode Standard for more
information about normalization (http://www.unicode.org/reports/tr15/).

For documented API that normalizes text, see IBM's ICU4J project:
http://oss.software.ibm.com/icu4j/