Held under the title “Bengali Language in Information and Communication Technology”, the seminar on the second day of SoftExpo 2017 brought together a panel of industrial professionals to discuss the problems in and the potential of using Bengali in digital platforms.
The 11th BASIS SoftExpo at Bangabandhu International Convention Centro had the entire length of the fair, from February 2 -4, packed with seminars and conferences in which digitization emerged as the overarching theme. This seminar being no exception, the speakers noted with importance a need for different actors in the field working together and advocated creating a space for Bengali across all platforms.
Saying that the government is designing a project to create at least 16 Bengali-based software tools, the speakers informed the audience that the project will cost around TK150 crore. This project may take at least two years of research and development, and the speakers urged individual developers and companies to work with the government to bring the project to fruition.
Mustafa Jabbar, the creator of the first Bengali typing software ‘Bijoy’ and a pioneering figure in the country’s ICT, talked about his major challenges when he first introduced the Bijoy keyboard. Even though the introduction of ‘Bijoy’ forged a new paradigm, at that period of time it was mired with compatibility issues of Bengali language in different platforms. Bengali was not usable as a general language on every system, as the necessary tools were absent at that stage for incorporating Bengali. With the advent of Unicode, however, in the late 90s, a new door of possibilities was opened and the process of Bengali computing began to take off.
Mustafa Jabbar started “Munir” keyboard, which was first developed in 1965 for use in typewriters, and the “ShahidLipi,” which is arguably the first complete Bengali typing interface. “I wanted to overcome many of the limitations associated with the existing options by developing a new Bengali interface with a new layout. My dream came true in December 1998 when the first edition of Bijoy was released,” he said.
Soon afterwards Mustafa Jabbar’scompany, Ananda Computers, successfully created a Windows-compatible Bijoy, which came in the market on 26 March, 1993. Improved versions were gradually developed to overcome shortcomings of the previous versions as well as to cope with newer versions of operating systems and hardware upgrades. Ananda Computers also developed a number of other software including interactive multimedia software for children’s education, a multilingual library management system and other training software.
Jabbar pointed out that machine translation for English has advanced a lot and English language sites dominate the web. But developing Bengalilanguage for machine translation and other necessary task have been and continue to be very challenging as developers face problems with font, machine translation, voice to text, text to voice, Bengali OCR (Optical Character Recognition), Bengali handwriting recognition system and Bengali spell checker. Though the IT specialists are working on all of these areas, Bengali is still quite far behind from English.
With the latest development of technology and increasing familiarity with English, people are not highly interested in using Bengali. Speakers also noted that there is no unity among the developers.
“Making Bengali software is a difficult task because there is such a shortage of people with the required expertise,” said SM Ashraful Islam, former executive director, Bangladesh Computer Council.
Besides, there are shortcomings of manpower. Even though women are participating in various fields more than before, they are still virtually absent in the technological world. Moreover, there is a lack of knowledge about technology.
Speakers said that people need to be enthusiastic in working with Bengali Language Processing in IT. Enamul Kabir, director of Bangladesh Computer Council said, “Classical approaches to Natural Language Process (NLP), along with empirical and statistical approaches are needed most. We need to have pure algorithm to make our language technology better.” Remarking that organized efforts in software and computer based content and software system localization in Bengali are not very visible in the country, Kabir urged that some basic standards for encoding the language must be developed before any content can be generated or any application developed.
“In the meantime, we have taken up a project that focus on the development of language technology. We are working with a Bengali corpus which will be helpful for spelling and grammar checking, speech reconstruction, speech generation, topic detection, message understanding and many other related topics,” he said.
Samira Zuberi Himika, the chairperson for BASIS Standing Committee on start-ups and the founder and managing director of Team Engine Ltd. said, “We had a dream to have enough Bengali Content online so that all people can have access to it for all purposes. We launched a project in 2011 or 2012, where OCR (Optical Character Recognition) can be applied from online. We can be benefitted immensely from digitizing the libraries.”
She said that it will be very helpful if thousands of books in the libraries are digitized. Himika said that her organisation started out with 30 members when they launched the project on OCR one and a half year ago.The Finance Minister Abul Maal Abdul Muhith, State Minister for ICT Zunaid Ahmed Palak and Mustafa Jabbar were present in the inauguration ceremony. Himika said that Team Engine have been able to carry on without any grants or funding.“I hope that we would not think of ourselves as competitors, rather make good relationship between ourselves to make our country digital,” she added.
When asked by a panelist if there is any language tech for handicapped people, Enamul Kabir, director of Bangladesh Computer Council, said, “Of course, we made software for disabled people with language technology. We utilized braille and used different technologiesto make software for disabled people. If people come forward to work with language technology, we all will be benefited”. He also added that in doing so,support from the government, language experts and developers will be crucial.
There have been positive developments in BLP (Bengali Language Processing) research. The main purpose of that is to be able to use Bengalion digital platforms so that ICT reach the general public and it can use and learn ICT in the native language. Morphological analysis of the Bengali language continues to be a significant challenge. Morphological analysis is very much essential for developing software framework for application level support, such as spell checker, OCR, TTS, grammar checker, MT etc.
Speakers also opined that we still do not have a large and representative lexicon of the Bengali language and the various lexicons currently in use do not contain a large number of ever expanding colloquial terms and proper nouns. Now, there is a need to build a larger and elaborate lexicon.
Most of the current Bengali language computing tools are primarily based on Microsoft Windows operating system. As all the open source operating has greatly improved and support Unicode fonts, focus needs to be shifted to cover these free platforms fully, speakers thought.
BASIS president Mustafa Jabbar ended the seminar reiterating and reminding everyone that the success for Bengali will only come if everyone understands that they need to work together for this common goal.
- “I hope to come back and work on nanotechnology in Bangladesh” - August 14, 2021
- ‘I believe stories of Bangladesh deserve to reach a global audience’ - July 31, 2021
- ‘Daraz is the synonym of customers’ digital lifestyle’ - July 24, 2021