Sridhar Vembu's Blog

Perils of Credentialism - MIT Example

Sridhar — Thu, 26 Apr 2007 18:48:50 +0000

From today’s news (courtesy WSJ):

The dean of admissions at the Massachusetts Institute of Technology resigned after admitting she misrepresented her academic degrees, the university announced today.

The dean, Marilee Jones, had worked at the university since 1979 and served as dean of admissions since 1998. She has been an outspoken advocate of reducing the stress of university admissions and had served on the board of the National Association for College Admission Counseling.

In a statement released by the university, Ms. Jones said, “I misrepresented my academic degrees when I first applied to MIT 28 years ago and did not have the courage to correct my resume when I applied for my current job or at any time since.

“I am deeply sorry for this and for disappointing so many in the MIT community and beyond who supported me, believed in me, and who have given me extraordinary opportunities.”

I don’t condone lying on the resume. Some punishment is warranted for the lying. With that stated up front, let’s think this through a little bit. For 28 years, she was doing a fine job. Then they find out that she falisified her academic degree in her original resume 28 years ago, and so she is let go ( “resigned” but presumably not very voluntarily). Does that means she wasn’t doing a good job all these years? But then if they ignore the lie and let her continue, wouldn’t that mean they are condoning lying on the resume? What is the solution to this conundrum?

Clearly, when she applied for the job, she felt confident enough that she could do the job, faked her resume, got the job. 28 years is sufficient time to evaluate someone’s ability to do the job, I assume. So what would I do in this situation, if I were the President of MIT?

First, she should pay a price for lying on the resume. I think the public humiliation of this incident being exposed is good enough punishment. I would let her keep her job. That would be a lesson in human fraility and a lesson in forgiveness to students, a bigger lesson than any number of moral lectures. Jesus and casting that first stone come to mind. A contrite and humble Dean of Admissions is anyday preferable to a suitably-credentialed, moralistic one. She would have genuine empathy for students who have different backgrounds, and different abilities. MIT would get a richer student body as a result.

As a university, I would go beyond letting her keep her job. To maintain intellectual honesty and consistency, MIT should announce that it would henceforth stop requiring formal credentials in evaluating candidates for this and other similar jobs. In other words, future candidates like her, who feel confident in their ability to perform the job, shouldn’t feel the need to invent degrees on their resumes. Come on, you may say, how are they supposed to find out who is a good candidate and who is bad. Well, they hired her based on an invented degree, didn’t they? Didn’t she work out OK for 28 years? Then why pretend that the degree was actually needed in order for her to perform her job?

Here is the reason to maintain that pretence: the university is in the degree granting business, so if they come out and say they don’t need formal credentials from applicants for jobs, they would be sending the wrong signal to their “customers” i.e prospective students. Yet, if they are actually intellectually honest, that is what they would do. Instead, predictably, they will convert this into a morality tale for students.

If you are a really smart MIT student, I encourage you to ask your administration deeper questions: “If we didn’t find out all these years, based on her job performance, that she didn’t have the degree she claimed she had, why does the degree matter suddenly now? What was the impact of her not having a degree on her job performance? Why shouldn’t students draw the real lesson that the university really wants to protect the economic value, to itself, of degrees it grants, but disguise that self-interested action under the cloak of a morality tale for students?” And be suspicious of faculty and administration who moralize about this issue, because there is an inherent conflict of interest for them here, and the easiest way to handle that conflict is to moralize.

The smart MIT student should realize, of course, that there are countless numbers of pioneers in this industry, and others, who don’t have degrees. They should give thanks to the capitalistic system, that sorts out people’s real merit in the market place, based on real job performance, rather than based on paper certificates that capture a very narrow slice of a person’s ability. Bill Gates could never have been the Dean of Admissions at Stanford, but the Computer Science building at Stanford carries his name. There is a lesson there somewhere. Don’t let the faculty brainwash you otherwise.

The “crime” of Voice-over-IP telephony

Sridhar — Mon, 12 Mar 2007 00:38:07 +0000

The police in Chennai seem to have a lot of time on their hands, for here comes the news Unauthorised telecom network busted, one held

Their crime?

Apart from Internet connection, the duo had also installed the voice over internet protocol with which they gained access to international calls, thereby causing revenue loss to the BSNL.

It was found that those who frequently made international calls dialled the local landline number furnished by …

Offering cheap international calls through voice over IP is a criminal offense in India. International calls in and out of India still carry really high rates. That segment remains one of the few highly regulated telecom market segments . In contrast, India has among the lowest mobile rates in the world, because of flourishing competition. Indian government could eliminate this “crime” in one stroke: simply legalize VOIP international telephony, and let competition do its job.

Credential requirements create a new caste hierarchy

Sridhar — Fri, 16 Feb 2007 23:36:48 +0000

I attended a short presentation on the book “No Asshole Rule” by Professor Robert Sutton. I have personally followed something similar to that rule.I believe anyone can be effective in their job, and be effective leaders without being assholes. One thing Professor Sutton mentioned was that the pervalence of assholes in the workplace was high in professions like medicine (arrogant doctors abusing nurses and other lesser breeds) and law (rainmaking partners abusing associates, paralegals and so on). It struck me how these are exactly the two professions where you need formal, state-sanctioned credentials to practice. Just about anyone can enter the software industry, and just about anyone can be the CEO of a public corporation - no formal credentials required. But even very experienced nurses, who are perfectly qualified to dispense medical advice, are prohibited from the practice of medicine. Of course, in practice, nurses do write prescriptions in the doctor’s name, but by law, they are told to be a lower caste needing the doctor’s name to sanctify what they are perfectly capable of doing themselves. All that the licensing laws have achieved is to create a state-sanctioned priesthood, and the inevitable caste hierarchy within these organizations.

Milton Friedman explained in the book Capitalism and Freedom how abolishing licensing/credentialing laws will improve medical care while lowering prices. The whole argument about unqualified quacks harming people is a redherring designed to keep the priesthood in power. In practice, chains and brands will emerge offering low cost, affordable medical care, and since they would have a reputation to protect, they will do the quality control on medical care.

On galvanometers and exams

admin — Tue, 19 Dec 2006 16:49:21 +0000

OK, answer this without consulting Google: Explain the principle of operation of the galvanometer.

That is part of the standard syllabus for our public exams in Physics for 12th standard students. I saw that question in a newspaper supplement, which provides sample question papers to prepare students for the all-important public exams.

I definitely should have studied this. In fact, I did well in the exam too. Except that I remember absolutely nothing. I could almost swear I had never heard of galvanometer, whatever it is, until I came across it in the sample question paper today. I am supposed to have an Electrical Engineering degree too, which makes it, well, more interesting.

Let’s agree that the purpose of education is teaching students how to learn, rather than fill their heads with facts and figures. Especially in the post-Google era, filling the head with such facts and figures does one absolutely no good.

Let’s also agree that the best way to learn how to learn is by actually learning something concrete. So that explains trying to teach galvanometer, which is a pretty concrete device. Except that it is impossible to explain a galvanometer (or object oriented programming or thermodynamics) in a high school setting well without really good faculty that understands these subjects, and such faculty is exceptionally rare. Galvanometer, which sounds like a fun device that I really have to learn about some day will become another water torture routine for kids. It would be a different matter if they were actually taught how to build such a device, use it to measure something, and so on. And it would help if they are not asked such dumb questions in exams, because then teachers would “teach to the test”, and no one will actually focus on learning how to build or use that device.

We are better off teaching kids bicycle repair, because that is a subject that can be taught well, and it also teaches kids how to learn. At least a question like “Explain how you would repair a bicycle” in an exam would be more obviously idiotic than a question like “Explain how a galvanometer works”, so at least it won’t get asked.

The point is not that everyone needs to know bicycle repair, but that they won’t remember anything about the galvanometer the way it is currently taught, while they will remember repairing a bicycle. And for those theoretically minded, bicycle repair can naturally lead to gear ratios and such. Of course the fastest way to kill any interest for kids in bicycles is to ask about gear ratios in exams.

Coming to the galvanometer, kids memorize something like (thanks Wikipedia!)

Quote:

A galvanometer is an instrument for detecting and measuring electric current. It is an electromechanical transducer that produces a rotary deflection, through a limited arc, in response to electric current flowing through its coil. The name galvanometer has been applied to devices used in measuring, recording, and positioning equipment.

That is merely an empty collection of words, with no particular feeling or experience associated with it, to a kid in school who is facing that all important exam. All that the kids get tested on is how well they can memorize that definition. I must have memorized something like that, and that was promptly flushed out of the system when those exams were mercifully over.

And the tragedy of it is that based on how they do in such exams, important decisions like which college, and what branch of study and so on get decided, and that in turn, plays a crucial role in the kind of jobs that are open.

Millions of kids in India are memorizing the galvanometer as you read this. I feel sorry for them. It is not just a monumental waste of time, it is much worse than that - kids that don’t do well in these exams are deemed unfit to hold any interesting job, because pretty much all such jobs will have stringent credential requirements.

Rising Inflation in India

admin — Fri, 20 Oct 2006 22:47:32 +0000

Here is a typical news story This Diwali is Costlier than Ever Before from The Hindu.

Along with inflation comes nonsensical “explanations”, from unnamed “economists” (emphasize mine):

Quote:

This Diwali, skyrocketing prices of essential commodities have left the “aam admi” [common man] very concerned. Be it sweets, dry fruits, crackers, pulses, dairy products, fruits or vegetables, the prices have registered massive increases compared to the Diwali last year. Even gold, the much sought after item during this festive season, is touching Rs.9,000 per 10 gm compared to Rs.7,000 last Diwali.

…

While in the case of pulses and food grains, forward trading in the wholesale market and the Government’s failure to check hoarders and big traders are blamed by economists, the steep increase in the price of diesel — a whopping Rs.6 per litre effected on three occasions between June last year and June this year — has also led to the overall increase in prices in a big way.

Gold has gone up nearly 30%, and essential commodities are up even more - Indians are gold-bugs, and I believe that is crucial to understanding the impressive stability of India in spite of all its problems.

The cause of all this inflation? Forward trading and “hoarding” are to be blamed, of course. Does it ever occur to any of these “economists” quoted that the Reserve Bank of India has been on a printing spree, fooled by the deluge of dollars pouring in to India, that deluge being Alan Greenspan’s gift to the world? And have these economists ever heard of the Friedman dictum “Inflation is always and everywhere a monetary phenomenon?”

This being India, the parliament and various state legislatures will soon pass laws to ban “hoarding”. They will unleash inspectors on traders, providing those inspectors ample opportunity to earn an extra income. Some essential commodities will disappear from sight altogether.

Inflation has consequences. The long run does arrive, before we are all dead.

On India’s Growth Story and Socialism Dying Hard

admin — Fri, 29 Sep 2006 22:03:23 +0000

NY Times reports on India’s Torrid Growth. Some excerpts (highlight mine):

Quote:

India is now growing faster than most other economies in the world, and is close to rivaling China, whose emergence as a manufacturing center has left India racing to catch up.

…

In a study published in May, the World Bank said, �It is easy to be optimistic about India�s economic prospects, but there is growing concern that the basic institutions, organizations, and structures for public sector action are failing � especially for those at the bottom.�

…

Still, economists and analysts warned that by itself, business dynamism is not enough for the long term. The high growth rate could taper off if the liberalization campaign of the 1990�s is not followed by a second wave of policy changes to share wealth more equitably and make it easier to invest and do business in India, the World Bank report on India warned.

If those policy changes are slow in coming, it is in part because the Congress Party-led governing coalition is beholden to the communist parties whose votes it needs to muster a majority in parliament. The communists oppose most of the measures that investors are calling for, including sales of government stakes in businesses, relaxing labor laws and opening the banking and insurance industries to foreign takeovers and competition.

�Has the pace of reform slowed down a bit? Yeah, it probably has,� Franklin Lavin, the American Undersecretary of Commerce for international trade, said by telephone from Washington. �Frankly,� he added, �you find more hangover socialism in India than in China. You don�t find someone arguing against you on the basis of Marxism in China.�

That highlighted quote captures the paradox of India. There is a small, but influential minority in India that is wedded to a romantic version of socialism, the “fabian socialism” of 1920’s UK. I have some close friends in that camp, and for them growing inequality only validates their socialist theories. Accoring to them, the fact that India stagnated for 40+ years and went nowhere economically with socialism is because, well, “we never really practiced real socialism” (thank God!). Socialism got Indianized, and became a gentle, bureaucratic beast, causing stagnation, keeping people in poverty, but not inflicting any obvious extra misery, at least none that anyone could directly attribute to it. The fact that India never experienced the horrors of Stalinist purges nor a Cultural Revolution means that romantic illusions could be nurtured without the harsh dose of reality getting in the way.

Now, with the Reserve Bank of India fully committed to the Alan “Bubbles” Greenspan program of unlimited, easy liquidity - to sterilize the deluge of dollars coming India’s way - India’s growth, just as America’s, is extremely lopsided. Credit-fueled specualtion, recently finding home in the spectacular real estate boom in India, pays far better than any kind of productive economic activity. The result is rising inequality, which socialists seize on to have their “I told you so” moment. Result: real economic reform is stalled, and freedom is the loser.

Measuring Programmer Productivity

admin — Sat, 23 Sep 2006 16:03:15 +0000

I often get sales pitches from vendors offering to sell us tools to measure programmer productivity. Typically these tools tend to be plug-ins to various IDEs, version control systems and the like. They may get fancy, and assign different weight to comments, white space, and lines that contain just a bracket, like

{

or even detect duplicate code - to give low weightage to cut-and-paste jobs - but in the end, they are basically measuring lines of code.

Once you start measuring “output” that way, you will tend to get a lot of “output”. Whether that output has value is a different matter. This is a basic Heisenberg problem (you will get a lot of whatever it is that you are measuring!), and I believe there is no escaping it.

I came across a company that proudly displays the fact that they measure every key stroke of the programmmer, and keep a webcam focused on them all the time as they work. I bet very smart people love to work under that system, and they get a lot of “productivity” out of them too. I hear this is a thriving business, particularly in professional services companies, where the quest to measure everything in software development is reaching its absurd limits, so clients can have “perfect visibility” about their projects.

From experience, I have concluded that human beings and measurement systems don’t mix. How do you measure commitment? How do measure passion? How do you measure drive? Yet, those things matter a lot more in productivity than lines of code. It is not a technological problem, and it is not worth thinking about it as “OK, our current system is imperfect, we just need a more perfect measurement system”. We need to rethink the problem fundamentally, or as in Zen terminology, just unask the question of measurement.

So let me ask it directly: Why do we need a measurement system in the first place? Is there a substitute for (very imperfect) human judgement in this area?

Price of Storage, Thumper from Sun

admin — Sun, 30 Jul 2006 17:45:10 +0000

Raw disks cost ~$0.5 a GB, and sinking fast. A high-end redundant storage system with all the bells and whistles cost anywhere from $20-50 per usable GB, coming down much more slowly than the price of disks. Only about $2 of the gap could be attributed to redundancy and the cost of hardware needed to process the bits and bytes before delivering them to an application. The remaining costs are accounted by software.

Along comes Sun’s Thumper:

http://blogs.sun.com/roller/page/jonathan?entry=the_rise_of_the_general

It marries a general purpose OS with storage, and delivers the whole thing at $2.50 per GB - and that box packs 24 TB of storage! Sun is very interesting, again - as a company that uses Java technology extensively, we are very happy to see Sun’s resurgence. Way to go, Jonathan!

Seth Godin on the Luck Factor

admin — Thu, 27 Jul 2006 22:09:38 +0000

Seth Godin has run his classic post on Factor L in his blog. He has been an inspiration for me for a long time, and this post tells you why.

I just cannot resist quoting him:

Quote:

Why do some products and services succeed while others fail? Why do some ideas go viral, rocketing across the marketplace, while others wither and die in obscurity?

….

We’d all love to know the series of steps to follow, the calculus we need to run, the hoops we need to go through in order to launch a product or service that is guaranteed to succeed. It would save us a lot of angst and disappointment if we could put our hearts and souls into something with a reasonable expectation of success.

It used to be that way, of course. If you had a very good product and a fairly sophisticated ad campaign (and enough money to keep both of them going), you could buy attention and turn it into market share. Creating new products was largely about having the will (and the means) to get your market to sit up and notice.

Today, it feels like more of a crapshoot.

…

As our marketplaces have changed, our approaches haven’t. We still overinvest to ensure success. We still make sure we have just a few eggs, in just one basket, and then we watch that basket really closely. Big mistake.

We live in a world of fashion, not rational computation. A world where everything from brake linings and ball bearings to clothes and airlines is chosen for unpredictable reasons.

The way to grow in the future is to acknowledge how important luck is and to diversify your risk. Do that with lots of products, not just one or two. Cut your overhead so you have plenty of chips, ready for another spin of the roulette wheel.

Priceless advice.

Toll Free Lines Down

admin — Thu, 27 Jul 2006 00:00:59 +0000

Our service provider had an outage, and our toll-free lines are down. We are working with them to get this back up. If you have urgent support issues, please use the forums, and we will respond ASAP.

(Update): The cause turned out to be power black-outs due to the heat wave in California. It impacted one of our telecom gateways. Now service has been restored. Thank you for your patience.