The history of machine representation of a written language is as fascinating as the history of the language itself. In the last few decades, the developers who have the understanding of computer programming have realized that language representation in binary form requires a through knowledge about the structure and development of the language. It is not only a tedious, but also a resource hungry process particularly for a language like Bengali, which unlike English and French, has never became the language of commerce. But the Bengali-speakers have always craved to "Bengalify" the digital world. After all in terms of numbers, Bengali as a mother tongue is spoken by the sixth largest population.
When did we feel the need for using computers to write Bengali? The demand for input and output facilities in the Bengali script was felt when these wonderful machines came to be used for office automation. It was not easy in the beginning and the use of Bengali script on computers has been a long, laborious history. But now there are about a dozen of software, most of them are add-ons to the Windows platform and few of them full-fledged word processors mostly on DOS.
No one is sure when people buckled down to use computers for writing Bengali script. There might be individual attempts that went unnoticed. As far as we know, a Sweden-based Bangladeshi student was reported to have first used the computer to publish a newsletter in Bengali in 1977. But nothing more, for a long time since microcomputers came into being and that was all he did or he was learnt to have done.
Three years later in 1982, Bangladesh University of Engineering and Technology (BUET) started work in this direction, and kept cutting its teeth on the matter for about six years. The university went for a hardware-based solution. Instead of writing software in graphics mode, the university devised a low-cost microcomputer in July 1982, crafted to use the Bengali script. That was the first time Bengali characters appeared on the screen, but they did not look great. The characters on the screen were of fixed pitch. Further the success was thwarted by compounding problems of conjunct letters. The project then shifted its focus. It tried to segmentize the Bengali characters and the work was completed in 1983. But everything was aimed at dot-matrix printer and could not satisfy the publishing industry. The whole project remained at the research level and was ultimately was shelved. At the fag end the endeavor in 1988, Mahbubur Rahman, BUET's head of the Computer Science and Engineering Department, published a report on his work in a computer magazine.
Taking the advantage of a GUI operating system, things started moving on the Mac front. Work on the design of font and keyboard started in 1984. Saifuddoja Shahid began to create a font on Mac in 1983 and finished his work in 1984. Shahid designed a bitmapped font, with the help of Apple's Resource Editor, which only worked in MacWrite. At around the same time, Zafar Iqbal, currently head of the Computer Science Department of Shahjalal Science of Technology University, developed a system to process Bengali script on Macs sitting at CalTech in the USA.
On the home front, the National Mass Media Institute (NMMI) was also thinking of doing something from 1984. Funded by UNESCO in 1986, the institute signed an agreement with Beximco to develop a full-fledged word processor. The institute received two Macs, two image-writers (dot-matrix printers), the ShahidLipi software and a laser printer. But it did not go further.
The first Bengali laser font for Macintosh, Bankim, was developed by a Calcutta-based company called Rahul Commerce. Gautam Sen, a man from Bangladesh, was the creator of this font. But the font was not up to the printable quality. At this stage, Shahid was asked to create a font, but Bankim had reached Dhaka before he finished his work. The second laser font was developed by Mahmud Hossain, a flight engineer by profession. By then, Shahid finished designing his one; but as there was no keyboard driver, typing with this font was very slow. ShahidLipi, ultimately, failed to sell in the market mainly because the laser font for the system came late; it used bitmapped font in the beginning; it was expensive and it could only be used in MacWrite.
The year 1987 saw another font coming, MainulLipi; but that was not, according to its users, a quality font. This font was used in a publication called Anandapatra by Ananda Computers Ltd, an IT firm headed by Mustafa Jabbar. The magazine came out on May 16, 1987. The problem with the font was that it could only be used from PageMaker. Users of this font used MacWrite for keying in the text and then did the layout in PageMaker. Later by September 1987, there were some improvements made to the MainulLipi font; but basically it remained the same.
In December 1987, Mustafa Jabbar developed a font called Ananda and it began to be used in a Bengali-language newspaper called the Azad. This font was better than the earlier ones. But using the Jabbar Keyboard, which was developed by Mustafa Jabbar in 1982 closely following the typewriter's Munier keyboard layout, it was very difficult to type faster. Because the keyboard assigned four keys to a single key on the board, in four tiers, normal, shift, option (same as the control key in PC) and shift option.
People wanted to type faster and in a very simple way. And they could do it. Mustafa Jabbar designed the Bijoy keyboard layout, leading to subsequent development of the Bijoy interface system and fonts like Tanvi, Sunanda etc. The keyboard and the Bengali-language interface were released on December 16, 1988. The keyboard interface was developed using Mac's assembly language, with the help of Devendra Joshi, then working as programmer in the Apple distributor's office in Delhi.
The keyboard manipulation became simpler and faster, with the help of a link key
He also designed the Sunanda font, "a perfect font for the publishing industry," according to Jabbar. This font was used by the newspapers like the Azad and the Daily Sangbad. Fonts like Tanvi, Rinkey etc. began to come out in a market that demanded more typefaces, especially for publication. Basundhara interface on Mac came out in 1992.
All these Mac programs were add-ons, running on the GUI platform concurrently with other word processors and DTP programs. They all dealt only with font and keyboard. No one thought of developing a full-fledged word processor. Up to the '80s, the whole affair was Mac-based.
By then the table turned for the IBM-compatible personal computers. People started work on using Bengali on PCs. The first attempt was made by a now-defunct company called Compico, which tried to use hardware-based Bengali in 1987. But because of low-quality printing and lack of ease of use, the program failed to gain ground. Another IT firm called Computer Land went for customization of a multilingual word processor that was used in the USA around 1986. The firm changed the codepage for the Devanagari script of the word processor called Duangjan and started marketing it in 1988. A few copies of the software were sold, but soon after, it began to be copied heavily.
Two local computer firms -- Computer Solutions Ltd. (CSL) and Graphics Information System (GIS) -- tried the localization process developed in India. Although the fonts created using the technology produced better results, the whole affair was expensive; and CSL and GIS failed to have any market share.
Shamsul Haque Chowdhury, of Automation Engineers, worked in the same line. He started working in 1987 and could use Bengali on computers with the help of a hardware card and a program called Abaha. Cards designed by the Department of Electronics of India eased the work of using Indian languages on WordPerfect-like software. Abaha was release in November 1988. Immediately after Abaha, Unidev Computer Solutions tried to market Indian GIST-card-based technology. But they both failed in a Mac-dominated print industry. Beximco Computers also had a try, but the project was shelved within a very short time.
Subsequently evolved Onirban and Barna. Onirban, which was not a basic word processor in itself, was released in March 1990. It was developed in Pascal programming language.
Barna came out to be the first basic full-fledged word processer on DOS. Work on Barna started, according to RAA Abdullah, one of the developers of the program, in 1988 and the first release came out in 1990. Developed in C, the program worked on DOS graphics mode and offered three Bengali and three Roman typefaces. The most striking feature of the software was that all its menus were in Bengali script. In addition to three popular keyboard layouts, it also had an "Easy" keyboard layout, allowing users to type Bengali after an experimentation of one hour. Bengali characters were mapped onto the corresponding English keys.
Despite all its limitations, the endeavor was a success for office automation. The marketing of the program, in the initial stage, was in the hands of another company. Later Abdullah, along with the other author of Barna, M. Shahidul Islam, founded the SAfeworks which took over the marketing of the software. Barna by then was bundled with a spell-checker, Pandit, with 60,000 basic word entries. The spell-checker came out in 1993. The SAfeworks ported the program on Windows in 1995 and named it Barnana.
Around 1992, a Bangladeshi student, Maruf Hassan, in China was also learnt to have been developing a Bengali-script text-processing software using multi-byte codes; but nothing of it was heard thereafter.
All these DOS-based programs offered little facilities for desktop publication. But as they were good for office and personal use, they commanded the market for a long time.
With the advent of graphical user interface like Windows on IBM- compatible PCs, efforts on font and keyboard (driver and layout) got a great going. In 1993 alone, the industry saw several add-ons -- Bijoy, Basundhara and Barnana -- coming out. Onirban was upgraded to version 3.0 for Windows in the same year. Proshika Computer Systems brought out Proshika Shabda in 1994 and Lekhani (both on Windows and Mac) was out in December 1994. Another two programs -- Asha and Prabartan -- came out at the same time. After few years, in 1997, Hi-Tech Professionals developed Anmana. But each of them failed to gain the market.
All what was to be done was done by the middle of the '90s. It was a time for innovative ideas. In 1998, a mouse-clickable Bengali text- processing program, 8 Phalgun, developed by Microtek, came out, followed by Duranta Bangla in 1999. The latest addition to this repository is Natural Bangla, developed by CDS-IT.
Several IT firms -- Access Ltd, Micrologic, Flora Limited etc. -- brought out their version of the add-on, but none of them could make dent into the market already dominated by Bijoy, Lekhani, Proshika and Barnana.
Unix was an unexplored region for a long time. Although it is possible to process texts using Bengali script in Unix-based programs like LaTex and others, no one in this country has as yet tried to write a processor exclusively for this purpose. But there were many around the world who have been using Unix for Bengali-script text processing using various types of font, mainly developed by individuals or organizations for education and research. Many of these systems are freely downloadable from the Internet.
Now with a wide variety of programs on Windows platform, supporting half a dozen keyboard layouts -- Bijoy, Jatiya, Munier, Basundhara, Easy, Onirban, Shahid etc. -- and around 200 typefaces, Bijoy of Ananda Computers Ltd, still tops the list. Few of the programs are now even bundled with spell-checkers. Even then, there exist the problems of electronic data interchange (EDI). Almost all of the DOS and Windows software use their respective font coding standards, character sets and keyboards. A file created using one of these software cannot be edited with the other. They can be opened though in another computer, if the system used to read the file has the font file installed. And this is the major setback text processing in Bengali is still facing. Lack of a standard character set and definitive rules for text representation are playing behind this incompatibility.
This article was first published in
PC Quest Bangladesh February 2000 issue
Revised: 5 April 2011