In conversation with K P Rao – the brain behind the Indian languages’ keyboard

Have you ever typed in Kannada, Hindi or any other Indian languages on your computer or smartphone? If you remember, you need to type K for Ka, G for Ga and so on. Have you ever wondered who came up with this idea?

Professor Kinnikambala Padmanabha Rao or commonly known as K P Rao. Recently, I had the opportunity to interview this genius and I have presented the same below. But before proceeding, allow me to introduce him to you as briefly as possible. 

K P Rao was born on 29th February 1940 in Kinnikambala, a small village in Mangalore. In his long career, he has donned several roles such as scientist (BARC), engineer, (ECIL, Tata Press, Monotype), Professor (Manipal University and IIT Guwahati) and writer (Varnaka and other research papers and articles).

He has received quite a few awards for his achievements in the field of Science and Technology. The prominent ones are in 2013 when he was honoured with Nadoja and Karnataka Rajyotsava the latter being the second-highest civilian award conferred by the Karnataka State Government.

Read along to know more about K P Rao, his achievements, his latest book and how he designed and developed the keyboard layout.

Ready? Let’s go!

How did K P Rao design the keyboard layout?

Surabhi:

You are basically known for designing – the first one to design the Kannada keyboard layout. How did you come up with the design, what is the concept behind it and if there are any challenges that you faced during the process? Also, can you tell me about the software “Sediyapu”? 

K P Rao:

Yeah, of course, it’s the question. 

When I put this idea about and the way that it was implemented, I already knew all the problems of Indian language composition with computers. I had predicted whatever or had foreseen the kind of problems and the kind of solutions that were required. I had seen it all. So that is why it was possible.

This is going to be very detailed and descriptive. And also it is going to be partially technical. 

So, let me start from the technical portion of it first. 

Basically, “Sediyapu” was invented for or was put into use for using Indian scripts on computers apart from Kannada. Just the languages for the time being. And that it was implemented in Kannada is a different story. But it was putting those Indian language characters and Indian language scripts onto the computer screen. 

And I started on it because it didn’t happen with “Sediyapu”, it happened a few years earlier to it. As early as something like 1980 for that matter. When I was working in a typesetting company, a company which made type and used those kinds of things known as Monotype. 

They were selling some large computer machines, computer-based machines for setting type – putting the pages onto the screen. They were totally mechanical machines. And with those mechanical machines, it was finally possible to print the characters onto a photographic film and from that, it will come to a printed page – whether it is a newspaper or anything that you have. All that was meant for English only. 

Now, when such a machine was to be adopted or modified, some kind of changes were required for Indian languages. 

Please remember, all these were digital computers. So, we had already seen the digital computer business – like the question of defining the characters with the screen, the number of dots and bitmaps were already known. 

But the question was how exactly to put relation to the English language characters or English language positions for corresponding Indian language positions? 

We had to invent some very interesting methods. Basically, this is what the entire art of programming was at that time to be able to configure tools like this paystub software. This was known as the phonetic input of characters. 

In English at least, we have got 26 basic characters. It’s upper case characters, capitals, and then the numbers and all that. So the entire lot of characters that are required for printing an English page is just about let’s say less than 128. So the reason why it’s 128 is different, but it is less than that. That is all that you need for printing a page in English.

But when it comes to Indian languages, this is not sufficient. You need many more parts of character shapes from which you combine and put the Indian languages. Whether it is printing it onto the screen or printing as a positive thing finally on to film, the procedure involved is the same. And it involves exactly importing or putting all those complex characters in place. 

Just to give the proper example from Kannada or Telugu. The “Ka” character itself is printed in one particular way. That is just one character. And then we’ve got the “Ki” and “Ku” and “Ko” and so many other characters that are derived from that basic character. 

Yes, and all of them have got their own specific shapes, which is not really desirable in that sense. 

And then I understood the trouble is, suppose we have a word called “Akka” and then there is a lower subscribed or lower case characters which have to go under the basic character. Then it is going to be something very very different. You should have a totally new position. 

So it all becomes very very complex to put or generate all these characters and have all those characters. And to put them all together, finally to make a page or to show them on screen was not that easy. 

Adding to the problem, the characters at that time in all the computers were English only. Even now they’re still the English keyboard only. So just with these 26 upper cases, we had to generate all the characters that were required. And to do that was not all that easy. 

So we had to write a computer program that will accept the keystrokes from the keyboard and from that generate all the characters that are required in the final form. And possibly modify depending on how you input the characters, show them on the screen correctly the way that it will be printed, or at least very near to the way that it is printed. 

This is known as the phonetic input. That is what we invented at that time. 

The idea started from this. Though our names are not language-independent, the Indian language names are written in English characters only. With a kind of modification, we can read it.

For example. Can I say “Ki” or Surabhi if you say “Su”, “S” and “U” together you can read it as “Su”? Whether it’s upper case or lower case is different. 

But then I got a “Ku” to come “K” and “U” can make a “Ku”. So that part is simple. But then each language or each of the scripts that you have got shows them differently under the screen and to generate them you need something much more complex. 

Maybe you start with the S, and S comes on the screen. The moment you put it “U” afterwards, the first character may be deleted or overwritten or added on to a new kind of a character, which finally would appear as S U – “Su”. 

Got a character like “Ki”, now you have the K and that might produce Ka on the screen. The moment you put an I next to it, the K has to go and KI has to be generated in that place. Replacing and deleting entirely the earlier character that was generated. 

To do this was very complex. We can do it with many more characters. The final characters shapes and all – to generate them is a very complex thing. 

This is exactly what we did. We created and generated these kinds of things. And I wrote the programs that time with just the 26 keys that are there in the right combination and the right assignment would produce whatever characters that were required finally in that language. 

We also had an advantage. When we are talking about Hindi or Kannada or any other thing, we have got a few characters that do not have any meaning or assignment on their own. But only when it is connected to another character, gets meaning. 

For example, the “U” matra that you use after that “K” or a “Sa” makes it a “Ku” or a “Su” that by itself never occurs anywhere. So effectively that “U” when it comes at the beginning of a word that is just after the space in the line or anything of that kind, can produce a complete vowel, because the vowel “U” never comes within a word. 

So it is possible to generate and use the same U in two contexts. When you just test it independently it produces the vowel character “U”, which is very different from the matra character that it generates. But if it follows anything like a Vyanjana or a consonant S or Sa or Ka or Pa or anything, it produces a different character that has to be put in there to make that character look nice or to look correct. 

In the same way, I would produce a different character when it is independent or when it is the beginning of a word, and when it comes after a consonant. Obviously, it will modify the consonant shape into something very new as I told you earlier about the key generation. It removes this entire character from being new, so this comes into that picture. 

So to do all these, you need not think and do it. Automatically when you write, you write it that way. But once it is input into the keyboard, it was interpreted that way, understood that way, modified and put on the screen, right? 

Surabhi:

Yes, I remember this. I remember looking at my screen while typing in Kannada it becomes Ka and then it changes when you add U or I and others. 

K P Rao:

Yes, exactly, that is what I told you. But then the input is also something else which actually we thought when we finally made “Sediyapu”. That is something that. 

If you have to store the entire information of a particular page. Whatever characters that you have introduced, it doesn’t matter. After all, there are going to be computer files. It’s not important to know it by the shape of the character. But it is enough to know how exactly it is being input from the 26 keys that are there. KI for that matter, you need not remember it as a single character KI. It is the K and I that has to be remembered. 

It was this concept that lead to something like Unicode at a later time – doesn’t matter of all modifications that they did. They took a totally different turn at a later time. 

So it is possible to input stored characters and store the entire linguistic text irrespective of the script in a totally different way just with these kinds of characters that are there on the screen. This is what we implemented when we brought in the idea. 

It’s a very difficult thing. I know that it is technical, but I can’t help making it technical. 

Surabhi:

Yeah, anyone who would have typed or at least tried typing a little bit of Kannada even on mobile would grasp it. 

K P Rao:

I told you this was invented or implemented as early as 1980 and it has gone unchanged. Virtually the same kind of logic is used even today for every other Indian language. 

When we did that, we implemented this for Telugu for a newspaper Enadu in Andhra.  

They bought a machine from Monotype and for them, it was necessary to implement this kind of phonetic input. And that is when we invented it and put it there.

The second one to use the phonetic input was those machines that went all the way into Kerala. Matrubhoomi also used those kinds of things in Malayalam. Then some party in Chandigarh sort of used it for Hindi and all that. 

Kannada, they never used these kinds of things. Our machines were not sold to them – the Monotype machine. 

How did he develop “Sediyapu”?

Surabhi:

Can you tell me more about “Sediyapu”? 

K P Rao:

“Sediyapu” came at a much later time. Because I was very anxious to use it for Kannada also. I sort of reworked down that, modified it and made the “Sediyapu” so that Kannada could be used. 

Initially, it was meant only for the screen just to get dot-matrix out but showing it on the screen with a lower resolution at that particular time. And then again I modified it into high-resolution laser output also. 

To produce our Tulu lexicon, the entire Dictionary of Tulu – the six volumes were composed using characters made by me using software that was done with this. In fact, it was Tulu that came up first and then other Indian languages like Kannada or anything else. 

So “Sediyapu” in a sense was kind of an implementation trial if you want to call it for Kannada, which is actually the derivative of what we already had done for Telugu or Malayalam or everywhere else at that time. Much, much earlier – 10s of years earlier to that we had done all this work. 

Surabhi:

OK, just a curious question. Which programming language did you use back then?

K P Rao:

I used C most of the time. 

Surabhi:

Oh!

K P Rao:

Yeah, the initial version 2 actually. In fact, I used an earlier version. It went almost unchanged. C was just being introduced at that time, so we used C. Because it was the most convenient thing that was there and we had a book also. So that was available – The original C programming language edition. 

Surabhi:

Ok. By Dennis Ritchie?

K P Rao:

Yes, exactly. So we learnt, we taught ourselves C programming and we did that. Actually me and some of my colleagues together that time. It was free. It was used entirely even from the beginning. 

Surabhi:

Wow, OK. We already know how powerful C is. Anyways, can you tell me how did your journey begin in this direction? What is the main inspiration behind designing the keyboard or getting interested in this field?

K P Rao:

No, no, no I had a job to do – that is Monotype. They wanted me to sell their machines and the machines had to be sold. We had to have an input system. I had to develop it. If I didn’t do that, I wouldn’t have a job or I have nothing else to do.

So, to ensure I fill my belly I went ahead to invent something. It’s like inventing a knife to do a job that you want to cut. That is all. Nothing. No big inspiration in that. 

His love for languages and being a polyglot

Surabhi:

OK, so, I heard this. You speak quite a few languages. How many languages do you know? 

K P Rao:

Languages are secondary. I speak a few but that doesn’t matter. Scripts – I know virtually every other Indian language script including Urdu and all that. I wrote once in Urdu. But it’s not very far. In fact, the problem with Urdu is that it is not as phonetic as our other languages are. They have got some new sounds – a few less and a few more. And then, of course, going reverse or right to left is not a big deal. That is very fine. 

Surabhi:

So you know Sindhu Sarasvati – the Indus Valley script. You understand that also. You made it possible to print that language also. Can you tell me more about it? 

K P Rao:

Yeah, so that was on a very ancient machine. Which was not technically digital, or it was partly digital. And in fact, there was not much programming involved there. 

So if you had more characters, they will be just generated by a computer. At that time it was done by the CDC computer that they had at Tata Press at TIFR (Tata Institute of Fundamental Research). It was used for generating a book on Indian language script. 

The story starts with a learned man called Professor Iravatham Mahadevan. He had many doctorates and all that. He was the secretary in the Department of Industry in Government of India in Delhi and was given a Jawaharlal Nehru fellowship to do some work on Indus Valley script. 

Then he produced a table and a concordance for the Indus Valley script characters. This was sometime in about 1977 when the book was printed. He started working on that in something like 1972. 

So the data was prepared, the concordance was done, and then there was the job of printing it out. There was a semi-digital machine – I’ll still call it an analogue machine actually. It was known as Alphaset, a Chicago based machine that used visible forms or mechanical fonts of characters. But then they had to be printed that way in any book.

I could show you the characters that were being used. It’s very complex, with many complex looking characters. There were something like 380 characters or so and they had to be printed. Which had to be printed and where was decided by the TIFR computer. 

We didn’t have a job. Accepting that on all those character shapes had to be joined with hand sufficiently large, reduced in size, put in a proper grid position and mechanical grid. We had to expose that with the light known as phototypesetting. A flash lamp would expose the character onto photographic film and that would go next to each other the way that was desired. 

In fact, it was a very complex job. I had to draw all the characters with my own hand. I was doing all those kinds of things. Though I’m not an artist, I had to draw it and reduce it in size just to hide all my errors and then have it printed. 

Iravatham Mahadevan later resigned from his job. He died only in about 2018 at the age of about 88. He was born in 1930 actually. And then he was the founder of the source behind the Titan watches and all the Titan activities. More than that, Iravatham Madhavan was an ICS or IAS man, so he was very, very active. 

His interest in the Indus Valley script was because someone told him that the Indus Valley script might be a South Indian language. And Father Haras, one of the people at St. Xavier’s college was s history man. He decided that it could be Tamil.

And they had a sort of a character which was a fish, and so six fish would mean “aar mean”. So they had their own meaning associated with these things. And it was a very complex thing that he did. So he was interested. He wanted to try whether it is really a language? Does it convey any meaning and all that? 

Nowadays, the entire thing has been digitised – getting handled digitally. Rajesh Rao also worked sometime back – just about four or five years back. He thought about the possibilities of having these computers producing Indian languages, I mean producing these kinds of characters – whether it is a language at all was a question. They did some entropy experiments and found that the language is there. And there is immediately a response to say that no, it can’t be using their own data and understanding. So this controversy is going on. 

Well, this was the first time that Indus Valley script characters were printed large scale in a book or text form. After that time, maybe they will draw a few characters by hand here and there within some distributive pages, or they will take photographs of existing Indus Valley sheets. So this all happened in about 73 to 77 in that period actually. 

So that is the story of Indus Valley. It is the first Indian script to be printed by electronic means. 

Surabhi:

OK, so this was even before the normal Indian languages were printed – so this was the first one. 

K P Rao:

Yeah yeah, I told you then. The normal one that came into printing was the Enadu in Telugu with the phonetic input and all that again using laser comp. Monotype to laser comp machine which was made in Cambridge. It’s in fact the part of a Cambridge University PhD thesis by David Hedlund anyway. 

But then the machine was used and it came into existence for English. Very expensive but very fast and very accurate in that sense. So adopting that into Indian languages involved again a lot of digital work and the invention of the phonetic input. 

Just to make it possible finally. All this time I’m happy that I was involved in doing it, and that’s something. 

Is K P Rao’s latest book “Varnaka” really a novel?

Surabhi:

Wow. My next question is about your book. You recently released a novel called “Varnaka” in around I think September 2021 or something. 

K P Rao:

Yes yes yes.

Surabhi:

Yeah, I’m yet to read it. It’s on my reading list. So, can you tell me more about your book like what is the book about? 

K P Rao:

Yeah yeah, you know. Basically, I’m interested in language and grammar. Usually, whenever they speak about the Indian language and grammar, Chomsky, a great linguist, talks great about the components that Panini the first one. Some people say that Panini was a computer scientist and all computerization is possible from Panini and so many other things. They come out with circles. 

This is partially true. In fact, many methods actually look for something like a table lookup. But all that he did was he borrowed quite heavily from his earlier work. There were many mentors who worked earlier to Panini and had their own contributions to linguistics and language at that time. 

So I thought there’s no doubt that Panini did great work, but then it was based on lots of earlier work and possibly there’s not much later work to be done. But the most important part of Panini’s work that whatever standardisation he did for an experimental, a projected or a hypothetical language or developed language that is there was prescriptive rather than descriptive. You’re not trying to describe an existing language but you’re prescribing your language. Try to explain the rules for the creation of words and phonetic components and all that in his book. So that is what Panini did. 

What were the resources that were available to him? How did he do it? How was that possible was something which always sort of bothered me. So I thought I wanted to write in great detail about it. 

I found a few stories in Kathasarithsagara and so many else. Instead of the modern people talking about Chomsky and other things, there are a lot of people who are also talking about Panini. The last one who talked very seriously about it was, two or three people in 15th-century people – a Kerala person called Narayana Bhattathiri. Very well known for his work Narayaneeyam. 

So that part I thought I will go deep into that original process where it was created and what happened, how he brought it to people. How and what were the inputs? What were the kinds of things that we are looking for and all those kinds of things. 

See, there is a lot of analysis of language that has to be done which we call in our Indian language style Meemamsa. So there was some meemamsic work that was necessary. Some analysis of languages, analysis of the procedure, analysis of rules, dialectical sort of work on rules for grammar was essential. It had to be reconsidered and some parallels had to be drawn. Then all these had to be tabulated. Because a lot of people call it entertainment, lots of exposure to various languages or various dialects, various possibilities and various stories – stories about language before someone is initiated into a search engine. 

So I found two brothers of this kind. One of them was known as Varsha. I mean great Maharshi Varsha who actually, for some reason, ran away from home and wandered around the world from China to Egypt for all that I know. He learnt many languages and sort of was totally confused. So confused that he became virtually dumb. If you just give him a small job, a peanut, he wouldn’t have a word for that because he will have to search through the entire list of words that he has got and then that will take the time that you will call him dumb. So that is how things went with him. 

He had a younger brother who went in search of his older brother, who had run away from home. He went all the way to the foothills of the Himalayas, met someone called Sukeshaka Bhardwaj, a person mentioned in the Upanishads. He learnt a lot from him, had his lessons of all the meemamsas and then he also met two brothers – an older brother and a younger brother called Panini and Pingala. 

And there was a divine intervention where some powers come and say that these are the sounds that Dhamaruga makes or kind of a little drum makes, and that gave absolute moksha to one of them. And actually gave them all the rules that could be produced by a dakkan – when you sort of hit a drum, the kind of sound that would be produced. So they came with that and they produced whatever was essential at a later time. 

It was their story that I wanted to say with so many intermediate stories. Does the earth have a language? Does water have a language? Does air have a language? If there is a language without God? Can you get so fixed and still be sort of available for a language or amenable to a language? Can you call it divine at all?

When you don’t understand the language and you still understand things and also things I wanted to answer – that is the product of a huge thing. It ran into some 120 pages. 

Once it was produced, I knew that no one will understand each and everything. It is very hard. So I thought of some preludes just to include a little of a micro-history of a person like me in the village who are actually at the time of the story.

Anyway, I thought of the micro-history of myself for that matter, and then how a small person in a village gets exposed to the world and how knowledge accumulates. 

Finally, the book actually became K P Rao’s journey through language or everything in that case. It exposed to a micro-history of a small boy in a village, right up to whatever that kind of things happened. I enjoyed writing it. That’s all that I need to say.

Surabhi:

OK, why did you make it fiction and not like an article or series of articles or just a book? Why did you make it a fictional novel? 

K P Rao:

Yeah yeah. Even as a thesis there will not be any readers because even as a book I know that the portion that I’m actually interested in will not be read by people at all. 

Because there is no entertainment there. It is all hard knowledge. It looks like a linguistic review. 

So I thought I would just make it a part-novel. People who were serious and interested in still can visit that part of the book. So for whatever money they spend on or don’t spend, at least they will have some entertainment in the other part. That was the concept that I had. 

Yeah, see the one small unfortunate part was when the publisher wanted it to be named as a novel. I was not happy to call it a novel because it is not a novel as I said. But then the problem was, you know he said that to put it in a cupboard or a library, they should have a label that had to be there. So they called it a novel.

And that is unfortunate – it’s not a novel at all. OK, I don’t know. Technically it is a travelogue. It is my travel through the world of languages all the time. Yeah, right from the beginning to the end. 

Especially, I enjoyed writing the last portion of it called the 3rd part. Again it runs for 140 pages or so. That is an epilogue, which sort of rolls with that. 

I enjoyed writing it because it talks about whatever be the languages that are happening abroad and it’s failing. And no one has succeeded. But then the only success that got retained in describing the language wasn’t something. Or maybe just because those people were very, very adamant they didn’t want any change. I mean restricting their language very drastically. You cannot do so many things with that. But still, it’s related. You take this purity in that sense.

So that is what I just wanted to mention. You can’t write too many lines. There is the problem. It was already too thick and I didn’t want to add more pages.

And then there is the other thing also. You know, somewhere in between I said about the methodology that Panini has employed or adopted. It’s very interesting that it was more like a periodic table. Well, you can technically predict the things which are missing, like Mendeleev’s time. But unfortunately, people just called it a mathematical table – that everything is decided and you can’t make changes with it.

So only one who tried to make a change was our Narayana Bhattathiri. Who was shooed and finally, he went into oblivion in that particular sense. So it was a tribute to Panini and his method and also to the people who tried to make improvements and changes in their field. 

That is what the entire thing is about, actually. 

Surabhi:

OK, do you have any more plans to write other fiction like novels?

K P Rao:

Lots of them. No, no, not fiction. I don’t like fiction. This is not fiction. 

Surabhi:

OK, yeah, in the same way as this book? 

K P Rao:

Yeah, different thing. Working on a big thing. It is about when the language was being used as a weapon in so many arguments and everything. There is a lot of politics involved. I have shown it to a couple of friends and waiting for their response. 

He never stops learning

Surabhi:

So, my next question is, I think it’s obvious that reading is one of your main hobbies or interests. Are there any other interests that you have, apart from languages? 

K P Rao:

Thousands and thousands of them. See, the computers and internet archives have put anyone in total jeopardy like me. I don’t know. I can’t read more than about 400 pages a day. And then the amount that has been thrown around it is so much. 

And you know, I’m a very peculiar guy. I don’t like any strange rules. If I find a side door, I find something interesting, I go the other way and then possibly I may or may not come back to the main track. So that has been happening, and that’s why I get totally lost many times. 

I enjoy my life that is important. 

Surabhi:

Before we conclude, do you have any message for youngsters? Especially on researching? 

K P Rao:

No no no. One thing I’m sure, all good research has been done by non-academicians. The academicians have done nothing. Yeah, absolutely nothing. They only produce some junk publications, papers and thesis and it’s of no use and no insight on anything. 

Surabhi:

No, I am not talking about the research papers. Like in general, if a normal person wants to learn about something, how should they research about it?

K P Rao:

I go to the OCW library at MIT. I keep on taking courses for two months at least – one course or the other. I’ve very rarely only two courses I have given up. 

Well, there’s a course on lagrangians I thought I will do. Another I had to do for some personal reasons and then I got a message from the lady who was doing it. When Nancy Kochenower said why don’t you take that course that I’m doing? It’s a 26-hour course on the anatomy of the brain. 

Surabhi:

So, do you have any messages for youngsters? 

K P Rao:

I’m not a big man to say that. Live happily. It’s the only message that I got. 

Surabhi:

OK. Thank you so much for your time. I really learnt a lot. I enjoyed it. Yeah, thank you. 

K P Rao:

Yes, yes, thank you. Yeah, thank you for your interview. I hope you’ll find it useful. Bye.

Conclusion

So, that was it. 

Is your mind blown already? 

What did you learn from K P Rao’s journey and his thought process of problem-solving?

What is your takeaway from this interview? Let me know in the comments below.

Leave a Comment

Your email address will not be published. Required fields are marked *