Rediff Logo Infotech IBM Banner Find/Feedback/Site Index
July 3, 1998


HCL Infosystems

The Polyglot Plot The Polyglot Plot

Priya Ganapati in Bombay

Email this story to a friend. The lingua franca of the computer world remains English. But if computers were to suddenly become polyglots and speak every world language with equal fluency would it not lead to a stupendous growth in their use?

Would this not drive the super growth hardware and software businesses into a period of unimaginable hyper growth? What would be the implications of such a revolution on world economies and culture?

Fortune hunters in the Great Silicon Rush of the Nineties are not talking about it as much as they do about the World Wide Web. But then it may be too early in the game to let the cat out of the bag.

Yet the impending polyglot plot of the gigantic transnationals is much thicker in the air of India, that multi-racial, multi-lingual, workhorse of the digital domain.

Unfortunately, debates and arguments on the subcontinent are weaker for the lack of a detailed market study on localisation of computing in native languages.

There already are solutions in computing for those who would like to work in an Indian language of choice. Some work their fix through hardware like a fresh chipset on a card for the box or through software like a true type font library or Unicode.

But all fail to answer two distinct and equally important questions. This is an industry where products can conquer global markets or evaporate off the disk on the basis of a standard that either gains or loses popularity. So the first question is which particular technique of localisation would become the popular standard way of doing things? The second is more specific of India where the language of business is predominantly English. Why would most people pick up computer literacy before English literacy?

Yet, however incredible it may seem, Rediff On The Net's investigation uncovered that only one highly restricted market study on localisation was ever carried out in the country.

For starters, most government agencies like the Department of Official Languages that are supposed to promote localisation of software have no figures to go by when judging the demand for such software.

Except for the National Association of Software and Service Companies, none of the agencies or corporations involved in the business has studied the market.

Yet everyone is quick to justify his or her assumption on why localised software is about to become the order of the day.

The Manufacturers Association for Information Technology, the hardware industry's largest grouping, promotes localisation on the premise that the availability of localised software would lead to increased sales of PCs and peripherals.

MAIT Deputy Director Vinnie Mehta admits no study has been commissioned to determine the demand for localised software. Yet, he does not shy from calling localisation a "hot issue". "Localisation is very high on MAIT's agenda," he insists.

The basis for these assumptions is the perceived need for information technology products in local languages. Mehta explains "Internet is a repository of information. Search engines can be customised to translate the information on the Internet. When Indians realise that the Internet can be used to give information they will start using it in a big way."

Mehta is not the only one to admit that the localisation drive might just be based on presumptions.

Rediff's hunt for numbers landed it next at the doors of the National Centre for Software Technology. NCST Associate Director S P Mudur too threw up his hands. "I don't know if there has been a real market survey," he lamented.

Mudur and two NCST scientists are helping Microsoft in its localisation efforts by providing consultation on how Windows NT can be enabled for local languages.

Mudur too justifies the localisation drive. "In India, we have a 30 per cent literacy rate. Of this 5 per cent speak English and at the same time are computer literate. Even if we are able to capture another 5 per cent of the remaining 25 per cent, we are looking at doubling the market."

He points out to another feasible strategy: identifying the right type of information technology tools for use in local languages. "Intellectual ability is not directly proportional to the knowledge of English," Mudur reminds.

The localisation effort has to have a global standard so as to enable any software developer anywhere in the world to port his products on a multilingual platform. When the problem is considered on this scale, Microsoft, which makes Windows, the most popular line of operating systems, demands scrutiny.

To begin with, Microsoft too does not have any market data on localisation. Despite that, there is a massive effort within Microsoft to localise the just delayed Windows NT 5.0 in Hindi and Tamil!

Windows NT Marketing Manager Neeraj Chaabi confessed Microsoft has not conducted any study. It has only relied on reports from "popular government agencies and advice from the DoE (the Department of Electronics)".

"Today there are 1.3 to 1.5 PCs per 1,000 of the population in India. The key thing here is there are a lot of computer projects happening in the government. In the long term I would anticipate business in the 'B' and 'C' category locations like Kanpur or smaller places than that. Small businesses that are not able to do anything with computers because language is a barrier will then not have any problems," Chaabi elaborates.

But like it is always the case with Microsoft, not all agree with it. India's supercomputing superstar the Centre for Development of Advanced Computing has been burning the midnight tungsten on localisation since 1983.

C-DAC pioneered the hardware route to the localisation solution with the 'Graphics and Intelligence based Script Technology', or GIST. And despite the near unanimous opinion that hardware would lock in localisation and make it more expensive, C-DAC managed to reach the forefront of the market.

GIST Co-ordinator Chandrashekhar Raje acknowledges that C-DAC too does not have numbers on the market's potential but is quick to add "We have been in touch with most of the application developers and users who seriously wanted to use Indian language components on computers."

A few independent software developers too have jumped into the fray, hoping to cash in on the polyglot plot. But among them is one sole philanthropist, Harsh Kumar, a divisional engineer with the Konkan Railway. He wrote a true type font library based script for Hindi, Marathi, Gujarati and Gurumukhi. Then he did the next best thing. He began giving it away for free to thousands over the Web. Kumar would even mail people floppies at his own expense.

The evangelist claims "There is definitely a demand in small towns and cities. I have seen small-time sweetmeat vendors in Indore use computers. But, yes, I don't think any actual study has been done on this."

The only organisation that has done a study is the National Association of Software and Service Companies. The survey was done in September and covered the Hindi heartland of Uttar Pradesh, Punjab, Delhi and Madhya Pradesh, using a sample base of 100,000.

But NASSCOM Executive Director Dewang Mehta can only reveal that the "We found the results to be quite okay".

Another fact that Rediff has been able to pin down in this haze is that the government is the biggest customer for localised software products.

"At present a heavy push comes from Delhi for software in Hindi. Lot of the thrust is from central and state governments," Mudur concedes.

Chaabi agrees with Mudur. Harsh Kumar explains the phenomenon in greater detail. "Companies are not interested in individuals as most of their profit targets are met by orders from the government. Today (computer) ownership is high but density is low. This has to change. We have to work to make the market high in density so that there is increased usage of PCs in India."

Though government and semi-government organisations impose Hindi and local languages, it has not sparked off a trend in the private sector. The language of business is predominantly English and there is severe scepticism that software in local languages will take off.

Industry experts are quick to dismiss the misgivings. Dewang Mehta rates one of "our strengths" as the fact that "we are an English speaking nation". Once localised software is available, people will start using it in a big way is his intuition.

Vinnie Mehta agrees that the agenda is not to cut short or destabilise the existing system. "English is predominantly used in urban areas. But at the district level offices how many people use English?" he asks.

Mudur goes a step ahead and calls literacy in English a "transient phenomenon". "In office environments that are directly oriented towards babudom English is used. There are many other sectors where work is not done in English. After all a computer does not need a language of its own," he points out.

Raje is equally confident, if not more. "The sales figures have proved it. English is only used in the urban areas but computer usage is not restricted to these areas anymore," he says.

He elaborates that the service sectors should be more interested in having Indian language support with their systems like those that deliver electricity bills, railway tickets and insurance forms. "This is just the tip of the iceberg," Raje warns.

Once the need for localisation is unanimously accepted, the debate reaches the higher plain of technique. Hardware or software is the question that has been haunting the contenders all through.

Though C-DAC's hardware solution is more expensive and cumbersome to implement, it is very counterintuitive why the GIST package remains the market leader. The GIST card is a hardware add-on that must be installed in the PC by a technician.

The method has been roundly criticised by most industry experts. Sentiments run so strong that C-DAC is virtually blamed for killing the localisation process by giving it a hardware twist.

Raje returns the fire promptly. "C-DAC has perfected solutions as per applications. For data processing applications there are no alternatives. Specially on UNIX and other multi-user platforms."

C-DAC has a range of software products too which it claims is as capable as English based software.

But experts remain unconvinced. "Openness is lost with hardware. One of the major reasons why localised software has not taken off is that most of the people are not ready to make changes in hardware," fumes Mudur who is helping Microsoft develop its software solution.

Mudur then slips out his, and Redmond's, trump card. "Unicode," he flashes triumphantly.

Unicode, a 16-bit coding system, is a superset of the ASCII code. It has a unique encoding mechanism by which fonts in different languages can coexist. The repertoire of this code includes characters of the major scripts of the world. The Unicode Standard Version 2.0 contains 38,885 characters from the world's scripts like Russian, Arabic, Thai and Sanskrit among others.

"More and more developers are supporting Unicode. The Unicode consortium includes all the big names (Sun Microsystems, Oracle, Intel etc) in the business," Mudur assures.

He explains NCST's role in developing the localisation solutions within Windows NT 5.0: "Today there is no need for any special hardware within the operating system. There is a set of function calls. Enabling (Windows NT 5.0) means that text input-output would be done in the language of one's choice."

Chaabi seconds him. "Windows NT has a kernel or core operating system which gives the ability to treat localised software as a subsystem. Our approach is to enable the operating system itself by building a language subsystem inside it. Unicode therefore opens tremendous opportunities for all local application software. Windows NT will have Unicode with ISCII (Indian Standard Code) compliance."

He assures that all of Microsoft's future products will support Unicode instead of ASCII.

Kumar agrees Unicode would be revolutionary. But he is emphatic that we must try and develop alternate solutions until Unicode is incorporated in all PCs.

Kumar has critics. The most vehement, perhaps, is Mudur who denounces the font-based approach. "Under the circumstances nobody can do anything other than what has been done. But we should not get carried away and say that those are the solutions for us," he warns.

But Mudur's "theories" do not impress Kumar. "For some promise, that Microsoft is coming up with something, I cannot wait. My approach is that I shall try to do the best with what I have," he rebuts.

Kumar is full of interesting anecdotes. "There's a beggar who is hungry for three days and then gets a dried piece of roti. You can't tell him then to not eat that, as it has no nutrition. Only after he gets back some of his energy is he going to worry about whether the food's good or bad."

Closer home at Pune, Kumar has another detractor. C-DAC's Raje explains the problem with the font-based solutions: "Language implementation is not just font support. If one wants to support Indian languages in the true sense it requires a lot of developmental work. Encoding, decoding and display mechanisms are very important for language support and cannot be handled by this approach."

Mudur joins in with "Font-based approach also means that utilities like dictionary, thesaurus and find and replace functions will not be available."

After crossing the bridge from the necessity of localisation to how it can be done, the next obvious step would be in counting the chickens. And because the moolah is directly proportional to the market size, Rediff went on to look for estimates on that.

Again the industry has no estimates to work with. The excuse is patent: It is too early to think about sales figures or even talk about of the market size.

Vinnie Mehta refused to assess the market. "It's an intangible kind of thing. First you have to create the market. You have to get people to use this kind of software. You have to actually groom, develop and flavour this segment," he defends.

"There is a lack of literacy and IT literacy in India. You have to introduce people to computers and vernacular working. But definitely, the market has potential. It is a kind of chicken and egg story," he elaborates.

Mudur echoes: "We feel that there is a market but it does not exist now. It has yet to grow. It will grow the day people start seeking information on the Internet."

Microsoft has yet another long-term plan up its sleeve. Says Chaabi "We are looking at a long-term impact and thinking in terms of a more proactive commitment on the issue."

The few figures that Rediff could finally put a finger on were so diverse that little wisdom could be squeezed out of them. C-DAC, which has been selling localised software for the past decade feels that "if the software is distributed properly with the right price and with general awareness" the market could be worth Rs 200-250 million. Dewang Mehta is sure the market could easily touch Rs 3 billion. An estimate that is over a 1,000 times that of C-DAC!

So, does it all boil down to popularisation efforts.

"Once Microsoft comes out with its operating system we'll go on a very strong campaign. We'll have an awareness programme with celebrities to push the product," promises Dewang Mehta.

But MAIT is less patient. It has already gone ahead with Kumar's Hindi, Gujarati, Marathi and Gurmukhi fonts. They are distributing them in the railway engineer's tradition of freeware.

C-DAC is holding seminars, exhibitions and road shows. Raje claims they are spending over Rs 5 million annually to popularise localised software.

Microsoft is to join this popularising brigade only after the launch of Windows NT 5.0.

Collectively, all industry members whom Rediff polled could not put a timeframe on their efforts to spark the localisation revolution.

C-DAC keeps brandishing its 100,000 user base while claiming that the market is already there and they are the leaders. Microsoft does not seem to think so. It believes the market will gain momentum as soon as it releases the first of its localised software. It is confident that its localisation products would take off by the second half of next year.

"We believe that with the right kind of technology it will not take more than a couple of months for localised software to catch on with users," Chaabi boasts.

Did he say months? Vinnie Mehta predicts that localised software will hit "big time" in about 10 years.

"There has to be a framework within the government. Right now, India is fighting a more basic literacy problem. As of now, I cannot throw up figures, as things are very nascent. The market would take at least 10 years to bloom," Vinnie Mehta feels.

Dewang Mehta puts it much closer at three years. Initially, he refused to comment on MAIT's timeframe but later tactfully replied "In this kind of technology 10 years is a very long time."

Dewang Mehta's estimate is echoed by others in the industry. "By 2000, some standard products like word processors and spreadsheets would get localised. Products with a more horizontal base would be the first to become popular," Mudur predicts.

The standards controversy spills on to peripherals too. As of now there are five keyboard versions available in the market. Yet industry representatives are not ready to support any particular version. Thankfully, all agree that the issue needs to be resolved.

"MAIT does not support any standards. But I do agree that there is a need for a standard so that people do not go berserk with their inventions. DoE is working on this," was Vinnie Mehta's comment.

Kumar too agrees that there is an immediate necessity for standardisation. "Today, honestly speaking, there is no standard in the country. Till all the seven or eight agencies involved get together for standardisation of codes interchangeability between software is not possible. However nobody's standardised agenda should be imposed on another," he says.

Refusing to be drawn into the quicksand, Dewang Mehta nods "I think all approaches are required". Microsoft has decided to play safe too: "Our products will be compliant with the standards prevalent in the country. They will be pretty much like that of C-DAC's" Chaabi promises.

"It's a question of popularity, of what is broadly being used. Our intention is to do a standard compliant product," Chaabi justifies.

Many allege that C-DAC has used its clout to get its keyboard approved as the standard keyboard in the country. One enraged government official summed up the matter by saying, "A policy maker cannot be a businessman. He would always try to see to that his products become the most popular."

Raje denies that C-DAC is involved with the Bureau of Indian Standards. But in the same breath he agrees that "C-DAC has been very aggressive for standards. We have provided all the relevant help to BIS to make this happen."

Kumar should be a happy man. His fonts require no special keyboards or templates. They can be used with any English keyboard.

But Mudur points to the inherent flaw in Kumar's effort towards making Indian languages intuitive on an English keyboard. "The very idea of intuitive keyboard layout is oriented towards the English language user. He may not be the user of a localised software and this would defeat the purpose of making localised fonts intuitive on English keyboards."

Microsoft and NCST have developed a keyboard that is similar to that of C-DAC.

The next question is obvious. How much money have these organisations spent on their efforts so far?

"Peanuts. Not beyond Rs 25 crore (Rs 250 million) in the whole country," answers Dewang Mehta.

If localisation is so close to striking gold, why then has so little money been spent on prospecting? The answer probably lies in Chaabi's assessment that the industry is perhaps not looking at the correct applications to work with.

"We need data processing capabilities and not just desktop publishing applications. Until today, efforts have been in terms of layering and word processing. So localised software has not really taken off in a big way," he points out.

C-DAC feels that the non-compatibility of data is another major reason. "Users are scared of using a particular brand as it may not have compatibility with another one. Hence popularisation suffers," Raje laments. He emphasises that more efforts should be put in by the information technology industry as a whole.

Dewang Mehta has an entirely different reason. He feels that it has become more fashionable to use English and it is not really in to use Indian languages on the computer.

Mudur sums it all up. "I should have access to the same tools in Hindi as in English, otherwise I will prefer to use English based software," he feels.

Tempers run high in the industry when it comes to the cost of localised software. Both Mudur and Kumar roundly condemn the fact that the cost of using a computer in local language should be more than that of an English one.

"It is unbelievable that in India a Hindi computer should cost Rs 5,000 more than an English one," says an enraged Kumar.

Talking of money, there are a few start-ups that are waiting to ride the localisation wave on the back of giants like Microsoft.

One of them is Bombay based Softplus International which has a 'Multilingual Notepad' supporting Hindi, Marathi and Gujarati fonts in the market. In two and a half months, they have notched up sales of 300 copies and are quite satisfied with the performance.

However, the company claims that with this product it is just testing the waters for its latest idea called the 'Executive Desk' which would be available in Hindi and Marathi. Sudhir Kadam, managing director, Softplus International, claims to have encountered absolutely no problems in marketing his software.

"The market is waiting for these kinds of products," he declares. Softplus has also developed a multilingual keypad, which, though not a software, can be used as an accessory.

For instance once you get to 'Microsoft Word' you could click on an icon, which would display a keyboard with the Indian font layout.

Sources told Rediff that this multilingual keypad is under certification by Microsoft, which is planning to use it. Chaabi, however, called this "pure speculation". He assured that "The localised products used by us is developed at Microsoft for Microsoft."

Mudur too is dismissive of Softplus's solutions. "A good solution is one which is built at the operating system level itself," he reiterates.

Vinnie Mehta has probably the last word on why software in Indian languages has still not caught on despite a decade of efforts.

"In India you have to be very cautious. Things take time. They happen, albeit slowly. But then it is not right to put all eggs in one basket. A little caution is always good," he concludes.

Tell us what you think of this story