Joshua Ferraro

Yes. Koha fully supports UTF-8. It is possible to use and search on Arabic, Hebrew, Chinese, Bengali, Turkish, etc. language characters in records (or any other UTF-8 characters). Here are some examples:

Chinese:

http://opac.liblime.com/cgi-bin/koha/opac-detail.pl?bib=23717

Hebrew:

http://opac.liblime.com/cgi-bin/koha/opac-detail.pl?bib=23787

Arabic:

http://opac.liblime.com/cgi-bin/koha/opac-detail.pl?bib=23750

It’s also possible to have utf-8 branch names / patron names, etc:

http://opac.liblime.com/cgi-bin/koha/opac-search.pl
(see the ‘location’ drop-down box).
There are some things that could use work though. Some languages, such as Arabic, are highly inflected and unlike English, there are cases where the articles (in English a, an, the) are combined with the base of the word. A good Arabic search engine would need to parse words like that, remove the inflected forms, and find records that have the same word bases. Of course, I’m drastically oversimplifying the process.

The current 2.2 series of Koha cannot handle that kind of search complexity without a serious investment in programming resources. This is considerably easier when using the Zebra plugin, but still would require a fairly complex setup and configuration. Of course, the Koha community would be thrilled if a library wanted to sponsor such development.

Another component that must be considered when dealing with the Arabic language is the right-to-left preference. Koha needs a right-to-left template option to accomodate that reading style.

About the Author

Joshua Ferraro currently serves as official Release Manager for the Koha project. He is also LibLime's President, Technology. He has over eight years of experience working with open-source technology. Over the past six years he has worked to promote open source in libraries. His library-related research interests include cross-database searching, integrated library automation systems, data mining and web spidering.