face to face - ias

11
Face to Face This section features conversations with personalities related to science, highlighting the factors and circumstances that guided them in making the career choice to be a scientist. From Education to E-governance V Rajaraman talks to Madhavan Mukund Professor Vaidyeswaran Rajaraman, the pioneer of computer science education in India is a man of myriad interests. From research and education to industrial automation and e-governance, these are but a few areas to which Prof. Rajaraman’s expert contributions extend. He is cred- ited with the establishment of the first academic programme in computer science in India at the Indian Institute of Technology, Kanpur in 1965. He was instrumental in establishing the Centre for the Development of Advanced Computing (CDAC) to design and develop supercomputers in India using parallel computing technology. He and his group are credited with developing parallel computers besides designing a real- time control system for Bhilai Steel Plant, the training modules for Tata Consultancy Services (TCS), and the computer science curriculum for All India Council for Technical Education (AICTE), the national council for technical education in India. The author of 23 textbooks, including the first on computer programming published in India titled Principles of Computer Programming, Computer Programming in FORTRAN 90 and 95, he continues to be a prolific author and enthusiastic teacher at 86. Madhavan Mukund (MM): You began your studies in Physics and then moved to Electrical Engineering. What brought you to Computer Science? V Rajaraman (VR): I grew up in Delhi. I wanted to study engineering, but in those days there were only two places in north India for this—Benares and Roorkee—and it was dicult to get a seat. I qualified for a scholarship and studied Physics from 1949 to 1952 at Delhi University but my primary interest remained engineering. There were a few institutions oering a post- graduate diploma in engineering after a BSc. The Indian Institute of Science (IISc) was one Vol.25, No.1, DOI: https://doi.org/10.1007/s12045-019-0927-9 RESONANCE | January 2020 141

Upload: others

Post on 20-May-2022

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Face to Face - ias

Face to Face

This section features conversations with personalities relatedto science, highlighting the factors and circumstances thatguided them in making the career choice to be a scientist.

From Education to E-governance∗

V Rajaraman talks to Madhavan Mukund

Professor Vaidyeswaran Rajaraman, the pioneer of computer science education in India is aman of myriad interests. From research and education to industrial automation and e-governance,these are but a few areas to which Prof. Rajaraman’s expert contributions extend. He is cred-ited with the establishment of the first academic programme in computer science in India at theIndian Institute of Technology, Kanpur in 1965.

He was instrumental in establishing the Centre for the Development of Advanced Computing(CDAC) to design and develop supercomputers in India using parallel computing technology.He and his group are credited with developing parallel computers besides designing a real-time control system for Bhilai Steel Plant, the training modules for Tata Consultancy Services(TCS), and the computer science curriculum for All India Council for Technical Education(AICTE), the national council for technical education in India.

The author of 23 textbooks, including the first on computer programming published in Indiatitled Principles of Computer Programming, Computer Programming in FORTRAN 90 and 95,he continues to be a prolific author and enthusiastic teacher at 86.

Madhavan Mukund (MM): You began your studies in Physics and then moved to ElectricalEngineering. What brought you to Computer Science?

V Rajaraman (VR): I grew up in Delhi. I wanted to study engineering, but in those days therewere only two places in north India for this—Benares and Roorkee—and it was difficult to geta seat. I qualified for a scholarship and studied Physics from 1949 to 1952 at Delhi Universitybut my primary interest remained engineering. There were a few institutions offering a post-graduate diploma in engineering after a BSc. The Indian Institute of Science (IISc) was one

∗Vol.25, No.1, DOI: https://doi.org/10.1007/s12045-019-0927-9

RESONANCE | January 2020 141

Page 2: Face to Face - ias

FACE-TO-FACE

of them, and I got admission to study electrical engineering there. After joining, I switched tocommunication engineering—the system was quite flexible.

Prof. V Rajaraman (left) talks to Prof. Madhavan Mukund (right).

MM: How was your experience at IISc?

VR: The diploma was a three-year course. In my final year, Prof. Vincent Rideout from theUniversity of Wisconsin-Madison visited IISc. He taught a course on control theory and setabout building an analog computer—with op-amps using tubes.

This computer was primarily used to solve differential equations. When Prof. Rideout left IISc,the system was 85% complete, but it could only handle linear differential equations. Also, theoutput of a computation had to be stabilized and viewed as a waveform on an oscilloscope butthe oscilloscope interface was not working.

MM: What did you plan to do after the diploma at IISc?

VR: One option was to join All India Radio as a technical assistant. In those days, the govern-ment had an Engineering Service, and I decided to write the qualifying exam. While studyingfor the exam, I decided to stay on in IISc and pursue research—the equivalent of an MSc in

142 RESONANCE | January 2020

Page 3: Face to Face - ias

FACE-TO-FACE

Engineering.

For this, I had to build something, and I decided to enhance Prof. Rideout’s analog com-puter. I built a new oscilloscope interface and added a unit to handle non-linear computations.Resources were scarce, and I had to cannibalize components from the U.S. war disposal equip-ment. Some specialized parts I borrowed from other departments—for instance, a photomul-tiplier tube was obtained from the Physics department. Some components were even sent byProf Rideout by post.

One important lesson I learned during this process was that a computer was useful only if therewas a community that could utilize it to solve their problems. I had to seek out faculty fromdifferent disciplines to find problems to solve using this computer.

MM: So did you qualify for the Engineering Service?

VR: Yes, I did pass the exam in 1957 and I was offered a job in the Engineering Service. Inparallel, I also qualified for a Government of India scholarship for overseas study. There were20 such scholarships across all disciplines — history, economics, . . . The scholarship could beused either to do a PhD or to do 1–1.5 years of practical training.

I went to MIT to join the Electrical Engineering Department. I travelled by ship for a monthand arrived 10 days late! Here I formally studied courses in Computer Science for the first time— assembly language programming, computer design, . . . . I worked on SHARE AssemblyProgram on the IBM 704. I wrote a program to compute Fourier transforms—it was verytedious, and I actually came to hate programming!

In 1959, I moved to Wisconsin-Madison to work with Prof. Rideout. I took more courses inComputer Science, alongside courses in Mathematics and Stochastic Processes. My PhD thesiswas on adaptive systems and analog simulation.

After my PhD, I was entitled to work for up to 18 months on a student visa. I taught for a yearat Wisconsin-Madison as an Assistant Professor of Statistics and decided to return to India in1962.

MM: Did you have a plan when you returned to India?

VR: No, I had no job! The IITs were new and I applied. At Bombay and Madras, I was toldto wait for the advertisement. Kanpur invited me to give a seminar. IIT Kanpur was still in atemporary location in the city. The campus was yet to be built. I had to wait some months fora committee to be set up to interview me to get a job, but eventually, I did get an offer.

Soon after I joined, IIT started moving to its new campus. First, the library moved, thenthe faculty. In 1963, IIT Kanpur acquired an IBM 1620, the first computer in India with a

RESONANCE | January 2020 143

Page 4: Face to Face - ias

FACE-TO-FACE

FORTRAN compiler. Three US experts came to set up the IBM 1620. Though I had noexperience with FORTRAN, I gravitated towards that group.

MM: How did you get into teaching programming?

VR: The IBM 1620 was initially used to teach programming to the faculty of IIT Kanpurand PhD students in the science departments. The Visitors’ Hostel in IIT Kanpur had justbeen commissioned, and there was a big outreach effort to teach programming to people fromoutside IIT Kanpur—scientists from CSIR labs, faculty from other universities, . . . . Thesewere typically 10-day courses with morning lectures and lab sessions in the afternoon. I beganby handling the lab sessions and eventually started taking lectures when the US experts left.

MM: How did Computer Science evolve at IIT Kanpur?

VR: When the US experts left IIT Kanpur in 1964, H K Kesavan joined from Waterloo asthe Head of the Electrical Engineering Department. Kesavan pushed me to focus on comput-ing. At the time, there was only a BTech programme and he wanted to start MTech and PhDprogrammes. I visited IISc and IIT Madras to recruit students.

Around the same time, I got an invitation to spend a year at UC Berkeley. Here I got to teach afull load of courses in computer science and also attend courses and expand my knowledge ofsoftware.

When I returned from Berkeley in 1966, the MTech programme started, with an option ofdoing Computer Science as part of an MTech in Electrical Engineering. Soon after this, thePhD programme in Computer Science began.

Around 1971–72, there was a move to have a separate Computer Science programme. TheSenate approved an interdisciplinary MTech course.

In 1976, it was decided to launch a BTech programme in Computer Science. The first batchwas admitted to this programme in 1978 and admissions closed at All India Rank 40, whichsilenced the skeptics! The first batch had 20 students, of whom 17 went abroad for higherstudies. Notable among them was Rajeev Motwani.

Interestingly, these programmes started before the department was created. Finally a separatedepartment of Computer Science and Engineering came into existence, housed in the samebuilding as the Computer Centre.

MM: You left IIT Kanpur soon after these pioneering efforts. Why?

VR: The early 1980s were very turbulent times at IIT Kanpur. Life was disrupted by strikesand other union activities. Classes never stopped, and the hostel messes never stopped running,

144 RESONANCE | January 2020

Page 5: Face to Face - ias

FACE-TO-FACE

but it was very hard work keeping things going. Managing the staff to keep the ComputerCentre running was another challenge.

Around this time, Satish Dhawan, then Director of IISc, invited me to move to IISc andstrengthen Computer Science there. At that time, Computer Science was part of the Schoolof Automation that was dominated by control systems.

I was getting a bit disenchanted with IIT Kanpur and decided to move. It took over a year topersuade IIT to let me go. By the time, Dhawan had retired and S Ramaseshan had taken overas Director of IISc.

MM: At IISc, you became the Chairman of the Computer Centre and started setting up a centrefor Computer Aided Design (CAD). What were some of the challenges you faced, and how didyou overcome them?

VR: When I joined IISc the computer centre had a DEC1090 with 512 K words (32 bit-word)main memory, 800MB disk and around a dozen display terminals. Besides the interactiveterminals, there was a card-reader and printer to run batch jobs and print the results. Thecomputer was housed in a portion of the Central Scientific Instrumentation Centre with verylittle space for expansion. There was no research activity in the centre as it was primarily aservice facility. I was surprised to see a long queue of students standing outside the computercentre to reserve half-hour terminal time. The centre worked for only 12 hours. My firstjob was to get the centre working 24 × 7 which I was able to do by persuading the staff tocome on night duty. I also started an operator training program with about a dozen traineeswith BSc degree who were given a stipend. After training, they got good employment in thefledgling computer field. I also started a programmer training course for BEs which was alsovery popular. There were also some hardware issues such as frequent disk failure due to poorfrequency control by the Electricity Board and also frequent power failures. These problemswere solved by installing a frequency controller and a captive generator. Even though someof the problems were alleviated, and there were no more queues, the capacity of the computerwas not adequate to meet the requirements of the researchers at IISc. A better computer wasneeded.

As I pointed out, there was hardly any research activity in the centre. I was concurrently aProfessor in the School of Automation to enable me to get research students. I had to getresearch grants for research as IISc had very little funds. I sent two proposals to the erstwhileDepartment of Electronics (DOE) of the Government of India. One was to set up a Computer-Aided Design Centre, and the other was to be one of the centres of the Knowledge-BasedComputer Systems (KBCS) Development Project of DOE. The CAD Centre aimed to developCAD software tools to design Very Large Scale Integrated (VLSI) Circuits and train engineers

RESONANCE | January 2020 145

Page 6: Face to Face - ias

FACE-TO-FACE

to design and use these tools. These tools were then classified as dual-use technology by theUS government and import of VLSI design tools by India was almost banned. Our role inthe KBCS project was to design and develop Parallel Computers. The CAD Centre was alsosupported by UNDP which paid the foreign exchange component for computers and expertsfrom abroad to give advanced courses and also stipends to CAD centre employees to be trainedabroad on development of tools. The grant was substantial (Rs. 88.8 Lakhs and US$ 168,450)for the period 1984–1991. As there was no space in the computer centre, DOE also gave abuilding grant to build the CAD Centre. Dr N. Seshagiri was the driving force in DOE to getgrants and negotiate with UNDP for both the CAD and KBCS project, and I am thankful tohim for recognizing IISc as a centre. The KBCS project was approved in 1987 with a grant ofRs. 62.5 Lakhs and US$ 150,000 from UNDP.

MM: Did the CAD centre achieve the objectives that you had when you started it?

VR: The CAD Centre was a great success. Besides training over a hundred engineers from in-dustries and research labs in VLSI design, we were also able to forge a long-term relationshipwith the Technical University of Delft in the Netherlands. The group at Delft led by ProfessorPatrick DeWilde had developed a VLSI design software NELSIS (Netherland Large Scale In-tegrated Circuit Software) that was improved in collaboration with our scientific officers andinstalled at IISc in a minicomputer of CAD Centre. We also gave the software to ITI Ltd.and ANURAG (DRDO Lab) for their use in designing ICs that they successfully did. Theyhad been denied LSI design software by the US export control. An active research group wasestablished at IISc.

MM: What were the outcomes of the KBCS project?

VR: KBCS project was the first one that initiated research in the computer centre. Scientificofficers recruited for this project in the centre also registered for MSc and PhD degrees in theComputer Science and Automation Department (The School of Automaton was renamed CSAin 1985). The project sent ten scientific officers for six months training each to various Uni-versities and research labs in the USA. A close collaboration was established with ProfessorArvind’s group at MIT. Seven high-level workshops on subjects such a dataflow computing,compiling for parallelism, programming distributed systems, to mention a few were held andattended by students and teachers of IISc and other educational institutions and research labs.Ten PhD students graduated writing theses on parallel computers. Two low-cost parallel com-puters were built.

MM: How did the idea for a supercomputer centre at IISc come about?

VR: The Platinum Jubilee of IISc was to be celebrated during 1983–84. Dr. S. Ramaseshan, the

146 RESONANCE | January 2020

Page 7: Face to Face - ias

FACE-TO-FACE

then Director of IISc wanted a special initiative to be taken to commemorate it. A committee heappointed with me as one of the members suggested that the computing facilities at IISc werewoefully inadequate and needed improvement. Many Universities abroad had supercomputers,and it was suggested that IISc should acquire one. After discussion, a proposal was sent to theDepartment of Science and Technology (DST) to support IISc to buy a supercomputer. Dr S.Varadarajan, the then Secretary of DST, was supportive. He, however, felt that the Ministryof Human Resource Development and the Department of Electronics should be associated asthe cost of the computer was expected to exceed US$10 million. It was also suggested that itshould be a National Supercomputing facility.

As we did not know the full implications of the infrastructure required, the cost of running sucha centre, and the policies for usage, the government suggested that a group visit similar centresabroad and hold discussions. A group led by Dr Ramaseshan, some institute faculty and HRDand DOE officials was constituted, and we visited during June 1984 the supercomputer centresat the Imperial College in London and CDC centre at Minneapolis. After this, there was someclarity. It was decided that the Ministry of HRD will fully fund the centre. It would be a projectin the seventh five-year plan starting in 1985.

On 31 July 1984, Dr Ramaseshan retired, and Prof. C N R Rao took over as the Director. Hefully supported the project. I wrote a proposal to the Ministry of HRD that emphasised the factthat a centre without associated research faculty and scientific officers would not thrive. Theproposal thus included the cost of computer and front-end computer and graphics computers(US$ 15 million), 20 faculty positions (inter-disciplinary), 20 scientific officers, Rs.6 Crores forbuilding and infrastructure, and Rs.6 Crores per year to maintain the centre. It was approved,and the process of selecting a supercomputer began.

MM: Those were difficult days to import high-end equipment, both in terms of financial con-straints and governmental regulations in India as well as export restrictions in USA. How didthese bottlenecks affect the plans?

VR: In 1984, only two countries manufactured supercomputers–USA and Japan. The avail-able computers were Cray XMP and CDC 205 from USA and NEC SX1 and a Fujitsu machinefrom Japan. As supercomputers were classified as dual-use technology, export clearance fromthe US government was required. We thought that Japanese machines would not need an ex-port licence but were mistaken. Japan did not want to sell a supercomputer to India withoutpermission from the US government. Export licence required Government to Government ne-gotiations. I was thus in constant touch with a joint secretary in the ministry of external affairswho was the Indian negotiator. A high-level committee of the Government of India approvedthe proposal and the budget request of IISc in October 1984. We then proceeded to evaluate

RESONANCE | January 2020 147

Page 8: Face to Face - ias

FACE-TO-FACE

supercomputers in the market and prepared a set of benchmark programs based on the require-ments of the Institute to be run on the prospective supercomputers. Simultaneously, we starteda four-storey building to house it. The building required special features such as uninterruptedpower supply of 1 Megawatt (UPS, plus generators), under the floor air-conditioning etc. Acommittee visited USA and Japan to hold discussions with vendors. Meanwhile, protractednegotiations for export clearance started between the governments of India and USA. We hadzeroed in on Cray YMP 132, ETA 10E of CDC and NEC SX1E of Japan as prospective super-computers to run the benchmarks in early 1988. A delegation from the US State Departmentwith about a dozen persons visited IISc, inspected the venue, held discussions with the faculty,and it looked like an export licence would be given. We then ran the benchmarks and CrayYMP132 was the clear winner. It needed Cyber 990 and VAX 8810 as front ends. An orderwas placed with Cray with the condition that the computer will be delivered by June 1989 witha margin of a few months in case the export licence was delayed. Cray applied for an exportlicence. To grant the licence, the US government put several conditions that included on-siteinspectors and not allowing visiting professors from east European countries to use the ma-chine. These conditions were not acceptable to IISc. The US government did not explicitlydeny an export clearance. However, Cray was not able to supply the machine by January 1991even though they had built the machine, and it was ready for shipment. The high-level commit-tee met again and decided to buy a vector mainframe Cyber 992E, two CDC 4360 front-endsfor 992, VAX 8810 and a cluster of 48 IBM 6000 RISC workstations connected by a fibrenetwork to work as a distributed parallel computer. Besides, these we also bought high-endgraphics workstations. The denial of Cray was a blessing in disguise as with the money saved,we went in for a parallel cluster, several high-end graphics workstations, a very powerful vec-tor main-frame, and a campus-wide fibre optic network for remote access to computers in thecentre.

MM: In parallel, CDAC came into existence. How significant was this for the supercomputercentre at IISc?

VR: Prime Minister’s Science Advisory Committee of which Prof. C N R Rao was Chairman,appointed a Committee with me as Chairman to suggest how India could become self-reliantin supercomputing. Our Committee suggested that India should build parallel computers ratherthan try to replicate the Cray model that was becoming obsolete. Dr Seshagiri who was amember of the committee was insistent that DOE should be the nodal agency to steer a centreto build such a machine. The committee suggested setting up of CDAC as a registered societyunder DOE with the mission to build a high-speed parallel computer. It was immediatelyapproved by PMSAC and subsequently by the government. IISc bought one of the first PARAMmachines. I was keen on having a cooperative arrangement with CDAC to develop application

148 RESONANCE | January 2020

Page 9: Face to Face - ias

FACE-TO-FACE

software in various areas of science and engineering. This did not fructify as CDAC was shortof manpower with appropriate qualifications and was more interested in hardware development.PARAM was thoroughly tested and we gave feedback about hardware and software problems.It was not used extensively by IISc students and faculty as application software packages werenot available.

MM: You oversaw the rebranding of the Computer Centre at IISc as the Supercomputing Ed-ucation and Research Centre (SERC). What were the significant aspects that changed with thistransition?

VR: As I pointed out earlier, the computer centre was a service facility with no faculty or re-search interest. Research in parallel computing started only after I joined and got a grant fromDOE. I felt strongly that a new area of computational science was emerging. Computer sim-ulation was becoming a new paradigm in science and engineering research besides theorisingand experimentation. It required scientists to work closely with persons who were trained incomputer science to develop new techniques and application software – supercomputer being atool. Doctoral students were required to work along with faculty. I also felt that scientists in ar-eas such as biology, physics, chemistry, and engineering disciplines who were computer-savvyneed to work with professional programmers and computer scientists to develop sophisticatedapplication software such as Gaussian 92, NASTRAN, Computational Fluid Dynamics Soft-ware, etc. I also felt that research in supercomputer architecture and systems software was to beinitiated. The supercomputer and the parallel computer systems we had would provide an idealinfrastructure to carry out the research. In an academic institution, departments doing researchare respected. A mere service facility, in my experience, slowly deteriorates, which was the sta-tus of the computer centre when I joined IISc. This reasoning was accepted by Prof. C.N.R. Raoand the Senate of the institute. In the proposal sent to the government, I had provided for 20faculty and 20 scientific staff (primarily system programmers and system administrators). In1988, the computer centre was renamed Supercomputer Education and Research Centre witha doctoral program. During 1990–1994, six assistant professors were recruited. As I wantedinter-disciplinary work to thrive in SERC, the appointments were jointly with the Centre forTheoretical Studies, Solid-State and Structural Chemistry Unit, and Computer Science andAutomation. Doctoral students also joined.

MM: Your name is perhaps best known to students through your books? How did you get intowriting textbooks?

VR: When we started teaching programming after the US experts left IIT Kanpur, there wereno textbooks, so we had to write them. The first textbooks were produced in the IIT Kanpurgraphics section.

RESONANCE | January 2020 149

Page 10: Face to Face - ias

FACE-TO-FACE

It was difficult to convince publishers that books written from India were viable. My first dealwas with Prentice-Hall, who agreed to print from a camera-ready copy with illustrations drawnby my wife Dharma. This was 1968 and was the first book published by Prentice-Hall in India.The book sold 3000 copies in 6 months. For the second edition, they agreed to use betterquality printing!

The first book was on FORTRAN, and revised versions are still in use. As the ComputerScience programme developed in IIT Kanpur to include an MTech and then a BTech course,the need for textbooks increased. I guess writing books became a kind of habit!

I collect material for my books from a variety of sources—journals, other books,...Sometimes,there are excellent books available from the US, but these are too verbose. One needs to cutthe material down from 700–800 pages to 300 pages. My last book was on Parallel Computingwith C Siva Ram Murthy.

I keep revising my books. Typically, I update them every 2–3 years.

MM: Have you tried online teaching? What do you think are the main challenges today incomputer science education?

VR: I did record some lectures for NPTEL about 4 years ago. It is a very different experiencelecturing to an empty studio.

The most pressing problem today is the lack of well-trained teachers. Unfortunately, there isno easy solution. Most teachers are not truly motivated and have come to the profession as alast resort.

Research and teaching must go together. To maintain the correct mental attitude, the disciplineof research is very important.

MM: It was wonderful talking to you. Thank you for sharing your thoughts.

VR: Thank you for spending a lot of your time with me.

Prof. V Rajaraman is Emeritus Professor, Supercomputer Education and Research Centre atthe Indian Institute of Science, Bengaluru. Email: [email protected]

Prof. M Mukund is Deputy Director and Dean of Studies, Chennai Mathematical Institute,Chennai. Email: [email protected]

150 RESONANCE | January 2020

Page 11: Face to Face - ias

FACE-TO-FACE

Special Postal Cover Commemorating the 25th Edition of ADCOM

Honoring the computer pioneers of the nation, whose contributions have been significant to thegrowth and adoption of computing and communications technologies for mainstream India, aspecial commemorative postal cover was released during the 25th Edition of Advanced Com-puting and Communications (ADCOM 2019) Conference organized by the Advanced Comput-ing and Communications Society. The special commemorative postal cover was released on 05September 2019 by Dr Charles Lobo, Chief Postmaster General, Karnataka circle. The postalcover is dedicated to the pioneers – Prof. V. Rajaraman of the Indian Institute of Science alongwith Dr S. Ramani of the Internet Hall of Fame, Dr N. Seshagiri, founding Director General ofNational Informatics Centre and Sh. N. Vittal, the then Secretary of Department of Electronics.

RESONANCE | January 2020 151