Sunday, April 10, 2005

The IITs won't be the same again

Tehelka, dated 9th April 2005, has a story on the changes likely to be brought to the IITs. This report is based on the minutes of the meeting of the standing committee of the IIT council, held on 5th March 2005, headed by the Chairman CNR Rao.

Let us look at one of the proposals being discussed:
Instead of the two-stage screening process, only one examination may be introduced from the year 2006-07. ... [T]he JEE should be made more simple and based on Class XII syllabi. Performance in the Board exam may be used as criteria for determining the eligibility. The directors of all the IITs would examine the possibility of making the examination more student-friendly giving due credit to the school system, and thereby enhance the credibility of the Board exam.
The two-stage screening process happened when the number of candidates writing the exam increased phenomenally in the late 1980s. I think this process was introduced from JEE 1988. Since the time JEE was introduced, until 1988, JEE followed its own syllabi and there were three hour examinations in Physics, Chemistry, Mathematics and English. The ranking was determined from the scores in these examinations and Board examination for Std. XII were not taken into account at all.

Since 1988, there is a first-stage screening process through a multiple choice exam, and those who qualify go through the second stage exam in Physics, Chemistry and Mathematics.

It is unclear what the standing committe's suggestion is with respect to making the exams student-friendly. The exams are being used to filter out students and choose a smaller number - on par with the number of seats available across the seven IITs. What use is there in making the exams simpler? One can argue that the exam pattern may have to change to find the really deserving students, than those who learn how to crack the problems but may not have the real aptitude.

It appears that the main reasaon for making changes in the examination system is because of the "mushrooming of coaching institutions." However, the suggestions will not result in curbing the coaching institutions. There are coaching institutions for IAS exams, GATE, SAT, GRE, GMAT, JEE, CAT and any other abbreviations you can throw. When something is coveted, like a seat in the IITs, as much, coaching institutions are inevitable. Starting up more IITs - another 10 over the next 5 years, and increasing the number of seats in the existing institutions should help. There is anyway a need for high quality engineers. The Central Government could convert some of the existing engineering colleges (like it did with Roorkee) into IITs, bring in more funds and improve on the infrastructure.

It appears that the IITs are also starting all sorts of crazy courses at the undergraduate level, such as B.Tech in pulp and paper technology (IIT Roorkee), B.Tech in environmental science and engineering (IIT Bombay) and a course in Earth sciences (IIT Kanpur). This seems quite stupid to me. IIT Madras had courses in Naval Architecture.

There is nothing new in Naval Architecture. It is a specialisation in Mechanical Engineering with perhaps a bit of Electronic instrumentation etc. It is enough to offer a few more interesting courses as electives. I am afraid courses in pulp and paper technology, environmental sciences etc. will only create unemployable students, who will soon take up jobs in information technology sector coding for Google, or take up an MBA, and start selling soaps and napkins for P & G.

I cannot remember anything more than 5 or 6 options in Cornell University undergraduate engineering program. Civil, Mechnical, Chemical, Electrical, Computer Science and Metullurgy. I guess most American Universities follow a similar model.

It is only our premier institutions who come up with crazy ideas.

IITs are also planning to attract talented professionals to work at IITs. Since they cannot match the salaries offered by the industries, IITs will be looking at offering certain other perks such as employment opportunities for their spouses, education for their children etc.

The same tehelka issue also carries an interview with SG Dhande, Director of IIT Kanpur, but that is a topic for another post.

Friday, April 08, 2005

Tata's partner Rover collapses in UK

Financial Times: Rover collapses as deal fails

MG Rover is a UK based car manufacturer.

Rover Company Ltd. started in 1904 as an automobile manufacturer. Over the last 100 years it has changed several hands and reached the hands of BMW, the German automobile company in 1994. With recession all around in Europe, BMW decided to dispose off Rover in piece-meal. However this became an emotive issue in UK as BMW planned to close down the manufacturing plant in Longbridge, UK.

It was then that another consortium called Phoenix Consortium, with the support of the UK Government, agreed to buy the Longbridge plant and license certain Rover brands from BMW for a mere UKP 10, in 2000.

Phoneix/MG Rover then entered into an agreement with Tatas to import Tata Indica to UK and sell it under the brand name of "City Rover". This project was a bit shaky in the end of 2004 when there were noises from Rover side that the quality of the cars were not that great. The expected number of pieces did not sell. Rover was also looking at a close cooperation with a Chinese automobile company called SAIC (Shanghai Automotive Industry Corporation). Ratan Tata, while in South Africa, made some statement to the effect that he will take a relook at the agreement with Rover, if they were getting closer to SAIC. But then both issued a joint press statement that all was well with their City Rover project.

But within four months of that statement, MG Rover has gone into receivership. Tata's cars won't sell anymore in UK, perhaps after this month.

But would the Tatas be interested in acquiring Rover's Longbridge plant? They can get it for real cheap now. They could look at using the plant to directly produce and market own branded cars (Tata Rover)?

Thursday, April 07, 2005

Nokia to set up mobile handset plant near Chennai

Nokia to set up handset plant near Chennai

India has over 50 million mobile users. It is only now that a manufacturing set up is coming up in India. This means that well over 60 million mobile handsets have been imported to India (possibly even 75 million) since India went mobile in the early 1990s.

Indians have been paying more for the mobile phones because of customs duties and the costs of import. This has resulted in a grey market transaction offering entry level new handsets at as low as Rs. 2,000.

Besides the 2,000 direct jobs and 6,000 indirect jobs promised, the real value of the plant will be the technical know-how that will spread to the top level electronic engineers the firm will hire from India.

India will need a variety of mobile devices, doing much more than handling GSM or CDMA calls. Better data handling, local language modifications, increased data storage, ability to play MP3 or other audio formats, reasonable degree of computing etc. will all be required. While most of the cutting edge technology will still happen elsewhere, there may be higher degrees of Indianised development in India. As a starter, the phones manufactured in India may be made Unicode character enabled and UTF-8 compatible. This will allow for optional font fownload allowing users to handle Indian language text messaging or content browsing.

That is the next phase of growth in the mobile space.

Monday, April 04, 2005

Dishnet to launch WiMax services

The Hindu: Dishnet to launch WiMax services

I think it was in year 2000 that Dishnet launched DSL services in India. I had returned from UK at the end of 2000 and was looking for something better than dial-up. Dishnet DSL was a strange set up. The connectivity to the home was not through the existing telehone wires. Landline telephony was still under monopolies BSNL and MTNL. So Dishnet set up a parallel infrastructure by planting a few DSLAMs here and there and stringing the standard twisted pair copper wires from there to the homes.

Despite this, it was great. The last mile connectivity was 64 kbps. The connection was always on. Every once in a while, almost every 15 days in fact, the wires will get cut for various reasons. We complain and they will come and fix it.

The initial set up charges were close to Rs. 30,000, non-refundable. I paid that! Then, for 1 GB throughput, we were charged Rs. 2,000 a month. This high cost structure went on for a while. At some point in time - I think in 2003 - the per month cost for consuming 1 GB came down to Rs. 1,000. The speed increased eventually to as much as 512 kbps. I continue to use this connection.

VSNL eventually acquired Dishnet's DSL business last year. The service levels have gone up considerably since then. There is no line fault (at least for me) and any small problems once in a while are well attended to. Almost immediately.

In the meantime, I switched my landline to Bharti's Touchtel (now called Airtel Landlines). The idea was that I could look at switching to their DSL connection. However a few experiments with their connection didn't help much. The quality and pricing were not comparable to Dishnet's (now Tata Indicom Broadband). So I continue to remain Tata customer for broadband.

Now BSNL has launched DataOne. It is unlikely that I will now get a BSNL line home (given that I struggled in closing the two lines I used to have with them!).


Now the news about Dishnet's wireless access. In the recent broadband policy document wi-max frequncies were to be de-notified so that anybody could use it to provide wireless Internet service. I am assuming that is what Sivasankaran's Dishnet Wireless is using. Again, by being a first mover, they are going to gain considerable market share.

Dishnet Wireless was focused in this area even when Dishnet DSL's business was being acquired by VSNL. Why weren't others interested as much in this business model?

Much more than DSL, wireless Internet delivery in India has higher growth potential. It calls for setting up tower infrastructure in a few places in the city.

Sify uses a different mechanism. They have line-of-sight wireless for adaptive broadband connectivity to their Iways and from there, are using cables for home connectivity. This doesn't look useful, given that cables in India are by and large of dubious quality. Last mile as wireless seems to be the most robust solution, and certainly competing with DSL.

Sivasankaran's Dishnet Wireless will not be technically as savvy and capable as Tata or Bharti, but they will roll out a working service, will offer aggressive pricing and will use low-end marketing techniques and acquire customers. Eventually some of the big guns may look at acquiring them at a good premium.

Why aren't others looking at a similar approach in rolling out infrastructure in different parts of India and hope to be acquired? It will generate wealth to entrepreneurs and also decent infrastructure to the country?


I was hoping that the new broadband policy will spawn several cable-tv-provider-like wireless Internet service providers in small towns. Get hold of a decent leased line or DSL connectivity or even satellite connectivity or a combination of these. Then set up wi-max towers locally and start offering at least always-on, low bandwidth internet connectivity in small towns.

Despite the hype, DataOne will take eons to move to small towns. They will always have equipment crunch. They will go through needless tender processes, corruption, delays, inefficiencies etc. before offering connections in every district in the country. One cannot be guaranteed decent service levels.

Providing a boradband service through DSL is quite different from providing GSM based mobile phone service. BSNL's mobile service survives because it depends purely on the centralised switch infrastructure, towers and making the sim cards available. Even here, BSNL always struggles in getting the sim cards out to the market. The cards are always in demand. They are constructing towers all the time, and that is one thing they will do well.

Providing DataOne service on the other hand is more complex. You need to roll out DSLAMs across the country in each exchange where you want to provide this service. You may have to replace the wires in some places. You have to provide equipment in every home. You will have to replace every faulty equipment (modem, router). There will be more customer calls to handle. There will be billing worries. DataOne, I think, will be a disaster.

Thus, there lies an opportunity for entrepreneurs, who can provide excellent customer service locally, to offer wi-max based Internet service across the country. Their business will get acquired eventually so money can be made for all the efforts.