We are happy to bring out long pending 2nd issue of Localization newsletter. We haven't seen any major activities/release in past month, big news being that Gujarati now joins the list of supported Indian languages in GNOME, the other languages are Bengali and Punjabi. Some activity has started for Oriya and Telugu in last few weeks. An Oriya font is being developed and Telugu KDE translations have started.
Surekha and Raghavan (surekha@servelots.com and raghavan@servelots.com) start as Sarai FLOSS fellows investigating and developing Indic tool-set necessary for web-applications built using server side Java, to provide a platform independent interaction in a browser based environment. ?CodeConverters is a Java tool for converting text between ISCII, UNICODE and WX (alphabetic) for DEVANAGARI and TELUGU scripts. These conversions can be extended to other Indian Scripts. Why do we need this Converter? In JRE 1.4.1, the Devanagari Character 'Nukta' when encountered in any text throws an exception. This is a bug in JRE 1.4.1. This tool handles the exception and consistently converts ISCII encoded text to Unicode and vice-versa. This converter tool is available at http://www.sarovar.org/projects/codeconverters.
Using the Java Input Method Framework, an IME for Devanagari (DevanagariIM_wx.jar), which uses keyboard layout based on 'WX' Notation, is available with the Java Unicode Editor at http://www.sarovar.org/projects/javaeditor
Rajesh Pradhan says he has started work on making Opentype Oriya fonts from the Akruti Oriya ttf fonts. He has already made a Unicode ones of it. Download http://www.indlinux.org/oriya/oriyafont.zip to try it out. More on Oriya localization is available at http://www.indlinux.org/wiki/index.php/OriyaLog
Support for Bengali/Bangla has been added to installer of the upcoming Fedora Core 2 distribution. According to Jeremy Katz, Hindi, Punjabi and Tamil are also sufficiently translated, but not enabled by default as no fonts for the scripts for those languages are shipping with Fedora Core 2. http://www.redhat.com/archives/fedora-trans-list/2004-April/msg00058.html
If you found typing in Indian languages using Inscript keyboard difficult and wished for a simpler roman like one, here is your saviour. Dr. G. Nagarjuna <nagarjun at hbcse.tifr.res.in> from HBSCE, TIFR, Mumbai announces availability of phonetic input methods for Indic, developed by project students from Indira Gandhi College of Engineering, New Mumbai. They can be downloaded from http://www.gnowledge.org/Data/Objects/kblayout/keymaps.phonetic.tar.gz . These keymaps are based om Xmodmap and settings are available for use with GNOME desktop. Additional help if necessary can be had from http://www.gnowledge.org/Data/Objects/kblayout
Dave.J.Holl has made available a Hindi inscript Layout for XKB, which has support for ZWNJ at http://www.smyrph.net/david/projects/inscript
Nirav Mehta is a businessman with a heart. The young, 24year old entrepreneur was inspired to take up the challenge of translating open source software to Gujarati after seeing the IndLinux Milan Hindi interface that was launched at GNUnify 2003 held in Pune last year.
"We use Linux in our operations and saw an opportunity to contribute to the community. Gujarati is my mother tongue, so I took up the task of doing Gujarati translations," says Mehta. The most notable part of the Gujarati localization effort is the efficiency, speed and cost-effectiveness with which the translations have been implemented.
Mehta's company, Magnet Web Publishing Pvt. Ltd is a 100 percent export oriented unit that employs 35 people and uses PHP extensively. Magnet hired seven interns from engineering colleges and this team worked full-time on the translations. The GNOME translation was done in a record time of one-and-half months and subsequently the team translated around 25,000 strings of Open Office.
The sharp businessman that he is, all this was achieved at a comparatively low cost of approximately Rs 28,000. Each intern was paid a stipend of Rs 1,000 and Magnet employed seven interns for a period of four months. A target-oriented task master, Mehta divided the interns into teams of two people each and set targets for each team. A blog was set up on the internal server to monitor the progress of the translations. Once the translations were done, Mehta selected the best two translators from the seven member team and assigned them the responsibility of testing, quality analysis and bug fixes.
One unique feature of the Gujarati translation effort (as compared to, say, the Hindi translation effort) is that all the translators were located in the same place. Mehta says that this had many advantages. "Communication levels were high, so problems were fixed faster. We did not have to send mail to a list and wait for the feedback. Above all, we had a lot of fun doing the translations."
The Gujarati team now plans to translate Mozilla, Evolution and other applications. Mehta is putting together a business plan and has identified over 30 business opportunities for Gujarati computing. Gujarati Linux distributions, training, Indian language applications for Linux, matrimonial sites in Gujarati, community portals, Unicode enabling of web sites, etc are some of the business opportunities that he has identified. He feels that the Gujarati Linux Distribution will find a market with end users, government and educational institutions, and companies based in non-metro cities. "In non-metros, people still use their mother tongue as the language of business, so that's a market that's not been addressed by IT vendors. Non Resident Indians are another market because they want their children to be familiar with the Gujarati language," says Mehta.
Mehta sums up his effort by saying that there is
a business opportunity in reaching the masses with the Gujarati Linux
distribution. "But this is more than a business. It also has a social
benefit attached to it!"
Nirav Mehta can be contacted at <nirav at
magnet-i.com>
Telugu is the considered the 2nd most widely
spoken language after Hindi in India, but ironically the least
localized. A Telugu Opentype font Pothana2000 is available under GNU
GPL. It works well in KDE, but needs further testing in GNOME.
Translation work is also to be taken up.
Current log of activity is at http://www.indlinux.org/telugu/
To volunteer join Telugu mailing list at http://lists.sourceforge.net/mailman/listinfo/indlinux-telugu
A new mailing list has been created for
co-ordination of efforts of volunteers. Please drop in and see if you
can volunteer and make Debian better! http://lists.alioth.debian.org/pipermail/debian-in-workers/2004-April/000000.html