Today marks the official launch of a new open source project. Utkarsh is an operating system based on Linux and localized in the Gujarati language, spoken by more than 5.5 million in India's Gujarat state and worldwide. Utkarsh (which means progress or rising high) version 0.1 is now in beta testing, and the team is bubbling with ideas for future growth. Recently Mayank Sharma spoke with the young Gujarati entrepreneur behind the project, Nirav Mehta.Mayank Sharma: Why Gujarati?
Nirav Mehta: Gujarati is my mother tongue. I was inspired by the launch of the IndLinux Hindi interface in February 2003. That was the time when I decided to bring Gujarati to computing. A team of seven engineering students joined me in the expedition to provide the benefits of IT to our community.
English understanding among Gujaratis is not something to write home about. Most of the people get basic education in Gujarati. The people who get educated in English too don't have a high command of the language. The language barrier becomes a bottleneck in bringing and sharing a lot of opportunities. Rather than aiming for English competence (which could be a distant dream) we are aiming to bring the benefits to the native language.
MS: How many strings have been translated?
NM: We have translated GNOME 2.6 core libraries. That's more than 16,000 strings. It covers the base GNOME system and many utilities and applications.
MS: Any plans for OpenOffice.org localization?
NM: We have already completed OpenOffice.org 1.1 translation. We are currently working on creating an install set and merging the translations back to the code base. The team is burning midnight oil now to have OpenOffice.org ready for the launch on the 27th.
MS: Could you break-down Utkarsh, technically?
NM: Utkarsh is based on Fedora Core 1, GNOME 2.4 (now moving to 2.6), and OpenOffice.org 1.1. We translated the GNOME glossary first, and then moved to the portable object (PO) files.
The glossary is a collection of commonly used words in the GUI. It was difficult to come up with appropriate Gujarati words, as many of the terms in the glossary are technical and there are no direct alternatives available. The PO files contain the actual strings used in application. One good thing about the localization effort in GNOME is that you don't have to go look out for strings in source code. The PO files contain the English version of the text and a placeholder to put the equivalent in your own language. We translated all the PO files in the GNOME desktop and developer libraries -- the core that creates the desktop and provides utility applications. These PO files are converted to machine object (MO) files, which are used by the application when you launch it.
The application language choice is driven by something called a "locale." Locale is the specification of the date and currency formats, month and day names, and collation order for a particular language. Thankfully for us, the majority of the locale work was done by the IndLinux project.
GNOME localization was more translation work. The real challenges started with OpenOffice.org and LiveCD creation. OpenOffice.org follows a long and complicated process of translation and integration. OO.o compilation is also a lengthy process; one problem can fail the whole operation.
MS: What fonts are you using?
NM: We started with Padma, a GPLed font from IndicTrans. Now we have developed our own font called Rekha. We are using Rekha for our distributions right now, and are looking forward to developing more fonts.
MS: How many people are involved? How did you get them together?
NM: The Utkarsh development team consists of seven people -- Ankit, Ankur, Atit, Bhavin, Kartik, Khushbu, and Sweta. I am the official head of the project. Vaishali did the creative work on the logo and the Web site.
I had been fooling around with localization for a few months and was pleased to meet Khushbu, an acquaintance from my hometown. Khushbu and five of her friends were looking for a project as part of their curriculum requirements. They were happy to work on Utkarsh. Kartik joined the team after about a month. He is also from the same town and was looking for an internship project.
The real work on the project started in January 2004.
MS: What unique challenges did you face during the process?
NM: The development team was a blank slate when they started on the project. Bits of Linux education in their engineering courses and hacking around got them started. The translation work was not too technical and the team picked up speed quickly.
The biggest challenge in localization is translation. It may seem an easy thing, but it can be downright daunting when you are stuck with a word you cannot find an appropriate translation for. It can get even trickier if people are coordinating the activities over a distance, so we were lucky that the entire team is working from a single place.
Apart from a lot of big and small difficulties that any groundbreaking work faces, we haven't had any major setbacks, thanks to close communication, strong quality assurance, and effective guidance.
MS: What elements in Linux made localization easier?
NM: If it were not for GNU/Linux, there would be no Utkarsh! The inherent freedom and access to code and knowledge allowed us all to leap forward. If we had waited for a proprietory operating systems to provide Gujarati solutions, we would probably have been married and settled down by the
time the solution arrived.
MS: Why do you think a country like India needs localized software? Will it make any difference?
NM: India needs localized software to move to the next decade. With a staggering population and insufficient infrastructure, localized software and services could be the bridge that fills the digital divide.
There are skeptics who say people will adapt to English. We believe people will not only adapt English, but will mix it with their own language. Their own language will change, as it has over the past so many years. Yet, localized software and resources are important to keep the language alive; local language newspapers would have died a natural death if this were not the case.
MS: How is the localization development scene in India?
NM: Localization is very much happening in India. IndLinux has been working on localizing Linux to Hindi for more than three years now. AnkurBangla is the project that works on Bengali language. There is IndicTrans doing Marathi, PunLinux doing Punjabi, and a few groups down south too. Even Microsoft has released Hindi versions of its Office suite recently, and there is strong interest from government and other groups. The next two years are going to be critical for local language computing in India.
MS: Is there a market for localized software?
NM: There is a market for localized software and allied services, including hardware provisioning. The local language solutions in India need to be provided as a bundle in most of the cases. As computers have not yet reached every corner of the country, there is untapped opportunity.
There is definitely a strong market for localized software, especially when it runs on a freedom-based operating system.
MS: Are there similar proprietary or open source solutions available in the local market?
NM: There are many companies that provide proprietory Indian language solutions. India was using local-language-enabled software in the print media for many years. There is no open source (or any other) solution available for the Gujarati language except Utkarsh.
MS: How do you plan to distribute Utkarsh? Do you have a single-CD version and a multiple-CD version? Will ISO images be available for download?
NM: The current plan is to have a downloadable version available on the site. This will enable Gujarati support on Red Hat-based systems too. We are also distributing LiveCDs at nominal charges to cover the costs. We still need to work out the distribution strategies, but we may have a full distro a few months down the line based on the growth of the project.
MS: What is the pricing policy of the various versions?
NM: The download from the site is free of cost. We are charging Rs. 200 (US $5) for sending the LiveCD version by courier.
MS: How do you plan to continue this project? Are you looking for volunteers?
NM: We are in this for the long term. We are working on business plans and growth possibilities. Utkarsh will be actively developed at MagNet, and we would certainly like to have others join in. We are thinking of working out a network of partners throughout the world who will help taking the benefits of IT to the average Gujarati. So yes, we have open hands for all types of win-win associations!
MS: Thank you, Nirav. Best of luck for the launch.
Mayank Sharma is a 21-year-old technology writer/developer from India. He does his bit to highlight and strengthen the localization efforts in India and is working on connecting FOSS with students and the education system.