One day in February, Farah Rahman visited the office of the bureaucrat who oversees the Indian government’s efforts to use computers and the Internet to better service citizens (e-governance it’s called).
The office is in Lutyens’ Delhi, the part of the Indian capital planned by British architect Edwin Lutyens, and Rahman met the officer a little after 6pm, by which time most officials in the e-government department had left. She found the officer sitting in darkness—there had been an electricity breakdown—behind a table loaded with files, books and documents.
Mahesh Jayachandra. Photo: Hemant Mishra / Mint
Rahman is a petite criminal lawyer from Minnesota whose family moved there from Hyderabad in the mid-1980s. She moved back to India to work with a scientist who had invented a normal 101-key keyboard that could handle Indian languages.
Here is what he had to say:
Yes, I met Mahesh a few months ago, and was impressed with him and felt the product was interesting. But from an investment criteria standpoint there are key gaps that need to be filled. A VC looks essentially at the following four factors while deciding whether or not to invest: team (who is going to execute on the vision); technology (what is unique/different/sustainable); traction (is there any validation ideally from potential customers, partners, advisors) and market (who has the pain, is that market large enough, and the pain point big enough to have the product or service be a “must have” rather than a “nice to have”).
Admittedly, there is a catch-22 that most entrepreneurs struggle with, which effectively has to do with the fact that one needs the money to get traction, and often needs the traction to attract capital. In this particular case, the entrepreneur has to create a business plan around the product, which is an extension of the four factors that I mentioned earlier, including sales and marketing strategy (direct/channel and branded vs. white label), target market analysis/research including detailed conversations with potential customers and partners, detailed competitive/alternative analysis and sustainable differentiation, business model/pricing, organization structure necessary to execute and the overall potential and exit for the company. In the US, often engineers and entrepreneurs connect with business school students, who often have real life operating experience, to help with the planning exercise. There could also be advisors or even angel investors from either industry or the public sector with whom the multi-lingual keyboard idea resonates. Those advisors and investors, especially ones with name recognition can lend significant credibility to this venture through their association.
Bottom line is that interesting product or technology by itself is only a 20% factor in a venture’s success. The remaining 70% is driven by execution and 10% is sheer luck. In Mahesh’s case, the 20% potentially is there with strong IP protection, but the other pieces need to be put in place.
The way she says it, the officer was disappointed that all she had to offer was a keyboard. He asked for a sample, inspected it, and thrust it back at her. “Keyboards? We have (local language) keyboards,” he said.
And then the officer deputed another officer to walk Rahman through what she describes as a shiny new e-governance centre.
The entrepreneur Rahman works with is Mahesh Jayachandra, a neurophysician who has both US and Indian patents for keyboards in Indic languages—a group of Indo-European languages including Sanskrit, most modern Indian languages, even some Asian languages. In a show of obvious symbolism, he plans to launch the keyboard, which was ready a few years back, today.
Behind the story of the Brahmi keyboard, as Jayachandra has termed his product, is the story of how a neurophysician who hopes to one day build an electrical model of the brain ended up inventing a keyboard. And the story of efforts by several scientists before him to solve a problem that, on the face of it, appears unsolvable.
The Qwerty keyboard most computers are equipped with has 101 keys. That’s more than adequate to represent the 26 letters of the English alphabet comprising five vowels and 21 consonants. Each letter is assigned a distinct key, which means users do not have to press multiple keys to type a letter.
Mahesh Jayachandra invented the Brahmi keyboard for regional Indian language users, which works by simply arranging the keys in a particular order, and assigning one key for matras and another for halants. Hemant Mishra / Mint
Indic languages have far more vowels and consonants. Hindi, for instance, has 11 vowels and 33 consonants. Then there are the squiggles called matras and halants in Hindi that lend an entirely different sound to the letter they are attached to. The matras are attached to vowels and the halants to consonants. Were each letter of an Indic language to be assigned a distinct key, the keyboard would need to have at least 1,500 keys.
While users may like to think of it in terms of keys, scientists think of it in terms of Unicode, an international standard applicable across languages where each letter, number or symbol is given a unique numeric value that works across various computing programs.
Apart from letters and numbers, anyone working on a local language keyboard in India also has to allow for special characters such as Om (Sanskrit originally but used in almost all Indian languages) and others that are specific to one language. Tamil has one, Marathi a few, and even Telugu has some that serve as small weighing units (they were probably created to weigh the fabled diamonds of Golconda).
India thought it had the problem licked when Mohan Tambe came up with the Inscript keyboard in the 1980s. Tambe came up with the idea for the keyboard at the Indian Institute of Technology (IIT), Kanpur, and perfected it at India’s then freshly minted Centre for Development of Advanced Computing, or C-DAC. Inscript, a phonetic keyboard, forms the basis of most local language keyboards currently in use in the country.
Such keyboards, though, require users to perform multiple operations before they can type a particular character. Given the level of complexity involved, it could take anyone between six and nine months to be trained in using an Inscript keyboard. That may not seem like much, and, indeed, one reason why there hasn’t been much progress in developing alternatives to the Inscript keyboard could simply be that many scientists and the government thought they had this particular problem sorted.
But there are some who think that the unavailability of a simpler Indic languages keyboard is one of the factors behind the digital divide. And the quest for an alternative has continued.
In 2004, researchers from IIT Bombay came up with the concept of a local language keyboard designed on the basis of the structure and usage of languages.
In 2007, C-DAC partnered with an Israeli company to create a LooKeys keyboard, where a camera on top of the monitor captured the position of the user’s fingers on the keyboard and read the corresponding letter.
In between, in 2006, HP Labs won a Wall Street Journal technology innovation award in the consumer category for a so-called gesture keyboard. Only, it wasn’t really a keyboard. One of the scientists behind it, Shekhar Borgaonkar, department director, HP Labs, says it was created more to address the issue of access than to facilitate typing (one reason why it isn’t a keyboard). The product is simply a tablet on which the vowels, consonants and numbers of the Indic language are arranged. To get a matra derivative, the user simply uses a pen to scribble the matra on top of the corresponding letter. Borgaonkar claims the product has 95% accuracy. This writer tried it out for a few minutes—he isn’t an expert in Hindi—and discovered that the accuracy was closer to 100%. But it isn’t a keyboard.
Still, HP hasn’t really pushed the technology because, Borgaonkar says, it doesn’t see much demand for it. The product, however, is available through a test marketing company with which HP has a partnership.
Jayachandra’s Brahmi keyboard is a real keyboard. And it works by simply arranging the keys in a particular order, assigning keys for matras and halants, and making use of the Shift key (much like English keyboards do). Nitu Thakur, at St Joseph’s Boys High School (also called the European School to differentiate it from another school of the same name) in Bangalore, was one of its first users. A teacher of Hindi, it took her no time at all to master the keyboard. And she tested it out by typing a question paper for a Hindi examination. “It’s just like typing in English,” she said.
Jayachandra has both US and India patents for the keyboard. The patents cover all Indic languages, so apart from several languages spoken in India, he can also look forward to producing keyboards in Sinhalese (spoken in Sri Lanka) should he wish to.
Jayachandra thinks no one will be interested in just a keyboard—a CD with the software and a keyboard-shaped sticker with both Hindi and English letters can convert any keyboard into a Brahmi keyboard in minutes. His company, called Brahmi Computing, is looking to launch a computer, also branded Brahmi.
The box will come bundled with a Linux-based Hindi operating system, and loads of free software utilities—math, painting, games, astronomy, the works.
Jayachandra is hoping to make enough money by licensing the technology or selling PCs to get back to his first love— electrophysiology. He wants to be the first to build a proper electrical model of the brain.
Now 47, Jayachandra is a dark wiry man with a jet-black beard and equally black mane of unruly hair that has been made to submit to some sort of order on the day this reporter meets him. He is dressed in black trousers, a white shirt, a black waistcoat and Oxfords— concessions to the meeting, he confesses.
Born in Bangalore to an academically and medically inclined family—his grandfather was a general physician—Jayachandra graduated from the Armed Forces Medical College in Pune in 1985. After spending some time in Bangalore at St Martha’s Hospital and St John’s Medical College, he went to the US to study further (he has an MD and a PhD from the SUNY Health Science Centre, New York). While there, he became interested in the brain, more specifically the cortical column.
Over the years, scientists have come to regard the cortical column as the key to the holy grail—building an artificial or electrical model of the brain.
The cortical column is just a group of neurons or brain cells located in the cerebral cortex. The cortex is the part of the brain that deals with memory, awareness, thought, consciousness, and language. The cortical column is of interest to scientists because probes can be inserted into it at a right angle. It is also of interest to them because the neurons there have almost similar receptive fields, or area where a stimulus will generate a response from the corresponding neuron. Put simply, the cortical column is like an even or uniform electrical conducting field.
Understanding the cortical column and mapping it is at the core of an ongoing effort to build a simulated brain, the Blue Brain Project, the result of a partnership between IBM and Swiss university Ecole Polytechnique Fédérale de Lausanne (EPFL). As the man behind this initiative, Henry Markham, project director at the Brain Mind Institute at EPFL, said in a 2005 article in New Scientist (the same year the initiative was launched): “It will be the first time humans will be able to observe the electrical code our brains use to represent the world, and to do so in real time…”
Markham added in the New Scientist article that he and his team hope to use the model to understand how certain malfunctions of the brain’s microcircuits could cause psychiatric disorders such as autism, schizophrenia and depression.
Jayachandra’s interest in the column arises from his belief that chips of the future could be modelled on it. “Think of the cortical column as a chip,” he says.
So, when he came back to India in 1999, Jayachandra, who had become interested in computing, built a rudimentary supercomputer and started studying the monkey cortical column.
The effort—building the supercomputer, not studying the monkey’s brain—resulted in a company, Peacock Solutions Pvt. Ltd, one of India’s earliest supercomputer companies. It was at the Peacock Solutions stall at IT.com, an annual IT industry jamboree in Bangalore, that Jayachandra experienced his moment of truth.
Mornings at IT.com, even today, are meant for business. On most afternoons, the exhibition is thrown open to the public—part of an effort to popularize computers and computing. Jayachandra had some interesting show-and-tells running at the Peacock stall. The audience was impressed but some left muttering that it would have been better in Kannada. Jayachandra thought so too.
For the past eight years, Jayachandra has been working on the Brahmi keyboard. He was ready with the product by 2006, but says he thought it would be best to launch it after he had the patents for it. In 2008, he received US patents 7414616 and 7420543 for the Indic language keyboard. In 2009, he received Indian patent number 230234 for the same. He hasn’t spent the intervening period building an organization because that isn’t what he does. There are a few people like Rahman who work with him. The money for developing the keyboard, much lower than $2 million—“How do you put a value on the effort and time?” Jayachandra says when asked to come up with a number—came from family, friends and IT projects Jayachandra did, in Bangalore, and in Minnesota, where he spent some time raising money, setting up an open-source and local language software company Kalibonca, and practising his rifle shooting (the good doctor is into guns).
There has been a smattering of interest, from investors, media and users in the product. In 2007, Kalibonca, named after 19th century London street herbalist Dr Bokanky’s miracle cure, the Kalibonca root from Madras, received a rash of publicity from Indian media for its Hindi keyboard.
In February, Jayachandra met Mohanjit Jolly, the head of the local arm of venture capital firm Draper Fisher Jurvetson. Jolly was initially enthusiastic and then, says Jayachandra, suggested that the doctor build a team and an organization around his idea. Jayachandra also says that Jolly suggested he get in touch with the Helion Fund, an Indian venture capital firm. Jolly didn’t respond to an email seeking comment. Jayachandra says he decided there would be enough time to meet venture capitalists after he had the product ready.
In 2008, Ashish Sinha who runs PluGGd.in, a site that lists and profiles start-ups, reviewed Jayachandra’s efforts positively. Sinha says he was impressed by the Brahmi keyboard, but adds that its success will really depend on Jayachandra’s efforts to forge partnerships with companies such as Microsoft or license the technology out to them. Sinha is a sort of weathervane in the start-up space in India, but even his write-up didn’t induce much interest in Jayachandra—except from one person who thought the doctor’s efforts to build a local language keyboard were an affront to earlier keyboards that had been in existence for at least three decades.
That’s a common reaction in the keyboard fraternity. Some of it may have to do with the lack of knowledge of Jayachandra’s patents and products. And some of it may have to do with the fact that some scientists consider the local language keyboard problem solved. In February, after the officer walking Rahman through the e-governance centre was finished, he told her why he thought her product wouldn’t work—the Inscript keyboard was there and worked well.
Rahman says the officer then seemed to wrestle with his thoughts till he found exactly what he was looking for. It was a piece of advice for Indians who went to the US and then came back with magic cures for everything—including local language computing issues.
“You should go back to the US,” he said.