My New Year’s Resolutions (for 2009)

It’s 31st December evening already—the time to make NYRs (New Year’s Resolutions)!!

When it comes to a “resolution”, I suppose I do have something else to say as well—by way of peer-reviewed published papers. For more information, search for the word “resolution” on this page. I am proud of those papers—regardless of whether foreign and Indian research/education establishments (e.g. COEP, or University of Pune, or IIT Bombay, or IISc Bangalore) have taken any cognizance of them or not. But regardless of their choices, personally, I can no longer think of the word “resolution” without thinking of those papers.

Those papers were very much a part of my PhD  thesis research, and they were written (and published) way back in 2005. Subsequently, I also gave my pre-submission seminar on 12 April 2007. And I have also submitted my final bound thesis on 12 Oct. 2007. And, still, till date, I do not have an iota of an idea as to when my final examination panel will at all be appointed (so that the thesis could be sent to them so that they could read it and decide whether they want to come to Pune to examine me in person or not)! Phew! Some people [gulps a lot of words]…

Anyway, coming back to my New Year’s Resolutions proper, here I go.

– Department of Exercise: (1) Walk regularly (at least five times a week). (2) Perform “Soorya Namaskar” daily. The quantum isn’t important; the regularity of it is. So, say, about twelve namaskars per day might be OK, but the thing is to see that there are no more than four offs per month.

– Department of Pranaayam and Yoga: Do the basic course of Sri^2 Ravishankar’s Art of Living, iff doing so is possible without giving up non-veg food. Many years ago, he was shown on TV asking people to direct their angst towards him. “Give me your anger, frustration,” he said, “so that you could release it.” Well, how about directing a little bit of fun towards you, Sri^2? I mean, what if someone really means it in a friendly way (i.e., without malice)? Or is it that your followers can’t let me learn your teachings if I approach them calling you “Sri^2″?

– Department of Research: (i) Start and finish writing down at least two full-length papers, say, from the earlier TBD list, but do not publish them until the PhD examination is conducted. (Whether I am passed or failed in the examination is not a concern here—not mine anyway. But whether immediate steps are urgently taken to conduct my over-over-due PhD examination or not, sure, is!)

– Department of Coding: Start and finish design and coding of three (preferably six) small VC++ projects. (Unnamed here; but small; each might involve just about 1K to 5K lines of code.) Three, if PDF documents also are to be written. Else, six projects.

– Department of Thinking on Physics: (i) Complete reading about history of CoV. Also complete reading of Prof. Taylor et al’s papers concerning teaching action principles. Simultaneously, complete jotting down basic points in refuting their arguments as needed. (ii) Complete reading and understanding van Dommelen’s book (and equivalents) in QM—whether other students are available and interested or not.

– Department of Learning: Break into nonlinearity—its general fundamentals, as well as its salient manifestations in engineering mechanics (fluids and solids). Finish at least five books, and get a sufficiently good idea of the field to the extent that a good overview lecture can be given to M Tech students by the year-end. … I have had enough of keeping nonlinearity on my the periphery of my mind. I knew I had to learn it some or the other day. The new year seems just right to begin understanding it.

– Department of General Reading and Writing: Read general references for and jot down the points (or preferably, the full article) on: (i) What principle might be at work when homeopathy works? Cover also the ancillary question (which I do raise, whether practitioners of homeopathy raise it or not): What are the limits of homeopathy? When might we expect it not to work? How about its ill side-effects in the process of finding the right remedy?

– Department of Music: (i) Get all my CDs on my PC. BTW, all of my music collection is un-stolen, and yet, I also don’t always necessarily seem to side with the stands that music companies seem to be taking. (Detailed discussion /clarification on this, later. (This is a “trailor”!!) Actually, there are some areas where my thoughts might come as unexpected to many people, esp. so if you support IPRs (the way I do) and regard music as a (commercial) property (the way I do). And still, I believe, it is possible, sometimes, in some minor contexts, to disagree with the specific positions that music corporates seem to take on this issue.) (ii) Create a database of my music collection. (iii) Compile a list of songs I like, as per the “raag”s on which they are based. (iv) Also, compile a list of songs that are beautiful but not based on any Indian classical “raag.”

– Department of Philosophy: Write one article (and write down detailed outlines for the rest) on topics like the following: (i) Free-will, innate ideas, and reincarnation (anecdotal “evidence” and more carefully gathered data of certain researchers). (ii) Locate and read at least five good books on the Indian theory of Karma (including original writings from Jaina and Budhist traditions). Objectively clarify issues concerning ethics, including free-will.

– Department of Society, Fellowship, Etc.: Begin addressing all holders of earned doctorates, not by the special title “Dr.” but by the more basic “Mr.” and “Ms./Mrs.” (or “Shri.” etc.), so as to help remind them that my own Ph.D. examination is over-over-due and still not yet conducted. Such usage includes, for example, Mr. Manmohan Singh, Mr. Sukhdeo Thorat, Mr. Raghnunath Mashelkar, Mr. Narendra Jadhav, Mr. Pandit Vidyasagar, Mr. Vikhe, Mr. Prakash Wani, Mr. Anil Sahasrabuddhe, Mr. S. R. Kajale, Mr. G. V. Parishwad, … Similarly, Mr. Sia Nemat-Nasser, Mr. Zhigang Suo, Mr. Biswajit Banerjee, … [and, similarly, Ms. Jayathi Murthy, Ms. Michelle Oyen, ...] … And so on and so forth… Thus, if you find me addressing a doctorate holding person by the title “Dr.” in the new year (or later), please consider it as my mistake, one arising out of habit, purely. This policy of mine will remain in effect until the time that my own PhD examination is in fact conducted. Note, whether I am passed or failed in the PhD examination is not at all relevant to the policy; my being asked to appear for the examination is. I will put an end to this new rule of mine on the day of my PhD examination. (And then only. Not even a day earlier.)

[Credits: The idea of calling some paragraph as "Department of XXXX" was shamelessly lifted from Mukul Sharma's "MindSport" column---that weekly column which he ran for some 20+ years in "Times of India," until recently. It is not, however, clear whether such usage of the word "Department" was of his own making or not.]

—–

I plan to update this blog with the progress (or the regress, or the oversight, as the case may be) of all my NYRs. Which, incidentally, itself is yet another resolution! OK. I need to make it specific. Let me: I will update the status as to whether I have been able to keep up on my resolutions or not, once in every three months or so.

Some loose-talk on the P and NP categories of problems

Do you know what the algorithmic complexity classes of P and NP mean? If not, then look up on the Internet (say, at Wikipedia). Better still, pick up Dennis Shasha’s book of the title: “Out of Their Minds” (It was a great holiday read about a decade back, and it remains so, even today.)

Given below is a very brief (and a very poorly written) outline of what the terms P and NP mean like.

Computer algorithms (and, by implication, the problems they solve) can be broadly classified into two categories: P, short for polynomial-time, and NP, short for non-deterministic polynomial-time. (Mathematicians do not know for sure anything about the classification relations between the two classes—whether they are mutually exclusive or not, whether they are co-extensive (or identical) or not, etc. … ) One way to approach understanding this classification scheme is the following.

When it comes to estimating computational costs of problems, two things help us characterize the situation: (i) the basic (or the “inherent” or the “intrinsic”) size of the computational problem itself, and (ii) the time required for solving that problem using a given algorithm. These two things are separate. Let us look at each one of them in some detail below.

Just to take an example, suppose your problem involves FEM (finite element method) as in computational solid mechanics, or FVM/FDM (finite volume or finite difference methods) as in CFD (computational fluid dynamics). FEM, FDM, FVM, etc. all ultimately translate into a matrix equation of the form [A]{x} = {b}. (Not so, directly, always, but certainly so for the static analysis case anyway.) Which means, they translate into a problem of simultaneous equations, really speaking. If there are N equations with N unknown variables, then the order of the square matrix [A] would also be N. If the mesh you use for your CFD modeling is made finer, the order of the [A] matrix, N, increases. Thus, one can say that N characterizes the basic size of the problem.

There is another consideration: Suppose we are given the particular values of the matrices [A] and {b}, and then are asked to solve the matrix equation [A] {x} = {b}. In short, we have to find what values the elements of the unknown column matrix {x} must possess so that the aforementioned matrix equation would be satisfied.

In finding the solution, we may use, for example, Gaussian elimination, or some other technique. All these techniques require certain multiplications and divisions to be carried out among the elements of [A] and {b}. Now, it is a well-known fact that the computer takes a much longer time (it consumes more number of chip-cycles) to do floating point multiplications and divisions, as compared to doing additions or subtractions. Additions and subtractions are extremely fast (even if done with double precision); multiplications and divisions (and powers and exponentials) are not. Hence, one important way to predict how long a particular solution algorithm would take (say, to solve the equation [A]{x} = {b}) is to ask: How many steps involving multiplication/division are carried out in executing that algorithm. You see, not all algorithms require equal number of multiplications/divisions. Thus, they differ in terms of how many simple steps each has.

It is useful to express the number of (multiplication/division) steps taken in terms of a measure called “Computational Complexity”, denoted as O(N). The capital letter “O” stands for the word “Order” (of complexity). In finding out the values of O(N) for different algorithms, we are not interested in the details of the exact number of steps; only in a rough estimate of the number of steps.

Now, the actual number of steps necessary to solve the problem scales as per the intrinsic size of the problem. Hence, it is useful to normalize the expression for the number of steps, with the size of the problem (N).

Thus, in general, we have an expression of the following form:

O(N) = f(N)

where f(N) is some function of N.

For instance, it is well known that the Gaussian elimination algorithm is numerically fairly stable (even if it is not as stable as so many people routinely suppose), and that its complexity scales as N^3. (The expression for its complexity also include certain lower-order terms like those involving N^2, but these not being dominant are ignored in getting to the *order* of the complexity.)

The aforementioned two factors (N and O(N)) together determine how long a given computational problem would take.

Now, it is easy to tell what the terms P and NP refer to. Well, kind of.

The whole idea is to examine the nature of the expression for f(N). Is it linear in N? That is, does it carry the form: f(N) = k N, where k is some algebraic constant which is independent of N? Even if not, is it at least a polynomial in N? Or is it a really bad expression involving something like e^N or N^N, etc?

If it is the former, then we call it an algorithm of polynomial complexity, shortly denoted as “P”.

On the other hand, if it involves some other relation like the exponential or the factorial relation, then, such a function increases far too rapidly with the increases in N, and so, the complexity (or the time to complete the computational task) becomes impracticably large.

[Here, it would be useful trying out a concrete example. Plot the graphs of N vs f(N) where f(N) is given by: (a) N, (b) e^N, (c) log N, (d) N log N, (e) N^2, (f) 2^N, and (g) N!, for N ranging from, say, 1 to 50. You will immediately appreciate the point that the range of numerical values for complexity can be very wide.]

Often times, it so happens that even if the algorithm to solve a particular class of problems is known, its corresponding f(N) expression is so bad that it is easy to see that the computer will never complete solving the problem using that particular algorithm. The predicted time for the algorithm to complete execution even on the fastest supercomputer can easily exceed billions of years of computer time.

So, in such cases, it’s a bit contradictory. We know the solution—but we don’t. To be more precise: We do know the algorithm, and yet, we also know that the only algorithm that is available cannot complete its run before the Sun collapses (which would happen after billions of years). (Interestingly, the algorithm can be proved to be correct—and yet, it cannot practically complete the run on a real computer.)

A few more points (though I know this blog-entry has by now already gotten into the “novel” mode.)

Sometimes, there are certain helpful mathematical relations embedded within the elements of these matrices—helpful, as judged from the viewpoint of getting solutions faster. These are the relations that appear out of the particular *numerical values* that the matrix elements happen to have—not just their relative *placements* in the matrices. For instance, if the matrix [A] is in the tridiagonal form, then it is possible to solve the equation [A]{x} = {b} in a time of the order of just N, not N^3. This algorithm is called TDMA (which is short for Tri-Diagonal Matrix Algorithm). (OK, it won’t be quite N; there would also be a constant factor appearing in front of N, but the point here is that the constant factor would be, presumably, relatively small, and hence, practically, not very relevant in estimating the order of the computational complexity.) So, if the matrix model is amenable to the TDM algorithm, it can execute very very very fast. (The computational time can come down from years to a few hours. Literally.)

Given this observation, naturally, people have spent some time in finding out if matrices can be “pre-condition”ed so that they can be put in a form that makes them amenable to a faster solution. (There is also another reason for preconditioning: they also try to precondition the numbers if doing so can ensure stability in the subsequent solution procedure.) The idea here is that the preconditioning operation itself would be computationally much lower-cost but that it would put the matrix in such a shape that a much faster running solution algorithm can then be applied to the preconditioned matrix as the second step. Thus, though preconditioning is an *additional* operation, overall, it can still save time.

Another important observation: Practically, it is often possible (or preferable) to settle for an approximate but fast method of solution, if the approximate answer is “close enough.” There is an important category of algorithms that do just this: They supply a series of solution each step of which gives an approximate solution; each step in the series is increasingly better refined. For instance, the multigrid and multipole algorithms exhibit precisely this nature. What determines if an approximate solution is “close enough”? Answer: The purpose behind carrying out the computation.

—–

Most of the above discussion was from the viewpoint of the computational engineer and  not of the computer scientist. Let me add a little bit from the second perspective before closing this piece.

The P and NP classes, as we saw above, are decided in reference to the question: Can you give a polynomial expression for the given algorithm or not. If you can, then the algorithm is in the P class; if not, then, roughly speaking, it would belong to the NP class. This is not at all exact, but it will do for our purposes here. The origin of the NP class rather lies in the automata and languages theory, not in numerical analysis. The two classes are not theoretically well-related to each other except from the viewpoint of the computational complexity angle alone. … Indeeed, whether they are related or not is a million dollar question, literally… The trouble is, no published proof tells whether or why a P-class algorithm can or cannot be formulated for that class of problems which is known as the NP class. For more information on the million dollar prize declared for this open problem, visit The Clay Mathematics Institute, http://www.claymath.org.)

Actually, the absence of a polynomial expression (so far) is just one of the interesting characteristics of NP. The other is: ease of solution verification. What the latter means is the following.

Let us take the traveling salesman problem as a prime example of the NP class of problems. The problem may be stated thus: Suppose a salesman has to visit each of, say, ten cities; suppose those cities are, say, Mumbai, Delhi, Calcutta, Chennai, Bangalore, Ahmedabad, Hyderabad, Pune, and Kanpur. He may visit the given cities in any order, for example, Mumbai first, then Chennai, then Ahmedabad, then Delhi … etc. Or, he may very well choose some other oder for scheduling his visits, say, Delhi first, then Bangalore, then Mumbai… etc. And, of course, there is fare to be paid for the travel in between any given pair of cities. The fare will in general differ from pair to pair. The question is: If the salesman has to cover all the cities and if he is to visit each city exactly once, then, what is the best possible order in which he should visit these cities so that his total traveling expenses are provably lowest possible. (You may also pose this problem in terms of the total distance traveled. The objective would be to minimize the total km for a tour that covers all the cities, visiting each city only once.)

In this example, N, the size of the problem, is 10.

An “easy” (or absolutely sure-shot) way to find the solution would be to make a table consisting of two columns. In the first column, we could make a list of each possible route (that goes through all the cities but visits each city only once). In the second column, we could enter the total fare/distance which would correspond to this particular choice of the route. Once such a table, enumerating all possible routes, is constructed, it would be an easy matter to pick up the route that has the least cost associated with it.

OK. So, what’s the point if the solution is so simple?

The point is this. Before you can decide which particular route would be the cheapest one, you first have to enumerate all the possible routes. And it is this  step—enumeration of the alternatives and finding out the cost for each possible option—which can easily become a computationally expensive operation.

For just 10 cities (N = 10), it is relatively easy to enumerate all possible routes, but that’s only because 10! is such a small number (comparatively speaking). But the factorial function grows very rapidly with N. (If you are a student of science or engineering, right away look up the Stirling approximation for N!) For instance, 5! is just 120, but 10! already is 3,628,800, and 20! is a mind-boggling 2,432,902,008,176,640,000 (i.e. approx. 2.4 billion billions). … When N approaches a value of thousands or tens of thousands (and in practice, we do run into far larger models), then enumeration of each possible alternative itself begins to go out of the reach of even the most powerful super-computer in the world (or all of them put together).

“OK, but why bother?” you may say. “Is it really practical? Does it really matter whether a salesman’s budget goes a little beyond what is absolutely optimal? After all, in the real world, salesmen do travel, and they do sell, and companies do make profits… Why do nit-picking about optimality?”

The answer to that question is: For an actually traveling salesman, minimizing of tour budget may not at all be an issue. Yet, this category of problems is important… If you build a mathematical model of certain practically important problems like choosing the best possible sequence of the forwarding stations for routing of a telephone call (or of Internet data traffic) then that mathematical model looks exactly like the traveling salesman problem. And, billions of dollars (literally) stand to be saved if you can find a method that will tell the optimum route in a fast manner (fast, as in real-time).

OK. So, the traveling salesman problem is practically important.

And, it falls into the NP category.

One funny thing with these NP problems is that if you already have a solution, it is very easy to verify that it indeed is a solution as compared to any other practically posed alternative. Verification (i.e. the one done by sampling) is very easy. But arriving at the solution itself is not. The latter takes too long, or is too pricey. After all, all that you have to do in verification is to pick up some other route arbitrarily, say at random, and verify that between the two given alternatives, the route which corresponds to the actual solution indeed has a lower cost associated with it. That is what verification is all about. So, verification is easy—provided the solution to be verified itself is known in the first place.

Another funny thing with these NP problems is the ease with which they can come up in applications. … Sometimes it looks as if the whole practical world is full of just NP type of problems, and that the P category is rather an exception than the norm. … Incidentally, this precisely is one reason why much creativity is needed even in such seemingly mundane tasks as relational database design. Anybody could establish relations among tables, and proceed to normalize them using certain set rules, but deciding how to abstract what kind of table from what real-world objects… Now, that part is difficult… It’s pretty much an art… And it will continue to be so, precisely because NP is so easy to run into… As another example, consider the problem of optimally populating the school time-table—assigning teachers and students to different class-rooms in different time-slots. This seems to be very simple problem, but it (I suppose) can be proved to be in NP. … The practical world, indeed, is full of NP situations.

All this was a background, really speaking, to get to know a bit about the P and NP classes of algorithms.

Next time, I will give a particularly attractive example of a, vaguely speaking, NP kind of situation. I have been planning to do so for quite some time by now. And the next time, the description won’t involve even this much of mathematics, I promise! (Though, it becomes very tough to write precisely. I began wanting to write a verbal but non-mathematical description; found that I could not; so, ended up making this piece a “loose-talk”. If you have some suggestions to make it less loose, esp. concerning the description of the NP class, then sure do drop a line… Top rankers of IIT Bombay (AIR 4 and all) are especially welcome to attempt writing (LOL!))

Strengthening of the (Proper) Institutions of the Indian State

Several recent news items have been disturbing:

The Defence Forces:
The court martial of a defence officer of the rank of a general for taking kickbacks in the supply of “daal” (lentils).
The suspected involvement of a Lt. Col. in home-grown terrorist activities.

The Police Forces:
The regular feature of “hafta” that common Indians inevitably face.
The failure of the police in nabbing even those criminals who operate across the states. For example, the sandal smugglers in the areas adjoining Karnataka, Kerala and Tamilnadu, the naxalites roaming free across Maharashtra, Chhattisgarh, AP, Orissa, etc.

The Judiciary:
Recent articles concerning impeachment or other recourses for bringing to line (or evicting from the judiciary) a High Court judge.
The number of pending cases and the projected time to complete them at the current rate.

And, something involving all—i.e., all arms of the government as well as common citizens: global terrorism.

Obviously, at the same time that India began to shine and Indians began to make it to the global Forbes list, there has also been ongoing a grave process of erosion of all the proper spheres of governence. If the Indian state is to be strong, these issues must be addressed with urgency, and the concerned institutions strengthened.

I have always been arguing for the lessening of government controls in economy. But it does not mean advocacy of either parallel governments or anarchy. (The former has de-facto been practised by the Naxalites in certain parts of the country for quite some time now—certainly also during the times when the BJP was in power at the Center. And as to anarchy, the suggestion towards establishment of police stations being run by political parties has also raised its head already—even if none takes it seriously.)

As the elections loom nearer, governments in mixed-economies tend to increase government spending. No country or party is exception to this rule. (For example, witness the actions of the recent governments in USA, UK, France, India, etc.) In particular, the supposed “right-wing” or the “moralizing” parties are not, either. BJP does exactly the same as does Congress. One would, therefore, expect a lot of expenditure to get announced over the days to come. … What can one do?

… Think about it… It, actually, is a good question—not just a rhetorical expression!

If it comes to me, I would advise the PM Manmohan Singh and the ruling coalition (and also state governments—irrespective of what party rules them) to sieze this moment and announce some really good expenditure plans that go towards strengthening of the proper institutions of government. (For the question of what constitutes proper functions of government, consult Ayn Rand’s writings.) It would make good sense politically as well as electorally.

I realize that I have to be extra careful here, simply because I am advocating increased expenditure. So, here we go…

Some don’ts. …

Firstly, Mr. Singh (and also chief ministers), do it decisively, definitely, but without foolish catch-phrases to go with it. In particular, do not try to describe it like “world-class government” etc. (the way Mr. Arjun Singh’s socialist friends recently tried to do, via showering of job-promotions under the name of “world-class universities”.) Please, do avoid that temptation.

Secondly, Mr. Singh, do not wrap your packages around with any socialistic language—flowery or otherwise. Just avoid this temptation too.

OK. I understand that politicians would want to have a punch about it. Sure, they can have it. For instance, don’t announce funds only for equipment and hardware resources for the police force. Go ahead, announce handsome salary hikes for the police personnel too. After all, our police does need incentive, and (I am being honest here) will actually also function better with increased pay. A proper code of morality would not regard the police personnel as objects of sacrifice.

Similarly, for the armed forces. But here, a little tact will be necessary. We will have to first tell the world that any increases on defence spending which arise because of the hikes in the salaries of our armed forces are both just and well highly well-deserved.

As to strengthening of the judiciary, frankly, I have no idea what precisely is needed. But I find it amazing that with all the explosion in the number of law graduates, we still should have to put up with a system of judiciary that cannot get or hire the sufficient number of judges… I find it amazing. What goes wrong? Where? I have no idea… Will someone please enlighten ordinary citizens like me?

(I also have some vague feelings about the necessity of repealing extraneous or superfluous laws. We in India have one of the largest written constitutions, and also a very long and unwieldy system of legal codes. We need to make it all leaner. For instance, we must strike out the word “socialistic” from the Preamble to the Constitution, fully restore property rights, and repeal many of the other “good-to-haves” that are not actually proper rights… But then, this whole matter of jurisprudence is too complex… It’s better if our lawyers and judges did something about it… I mean, we must push for constitutional reforms too, but legal experts will have to look into the aspect of what time will be most opportune for undertaking what reform when… One doesn’t want to end up having a worse legal code than the one we have already.)

In any case, I stand for strengthening of these three institutions. They do correspond to the proper functions of government. And, the time is opportune… What do you think, Mr. Manmohan Singh?

[Note added Dec. 5, 2008: May be, the time is not so opportune anymore.... The above piece was initially written on Nov. 24, 2008. It was revised and published on Dec. 5, 2008. The revision was limited to just adding a few sentences or so: the two examples of the difficult-to-pin-down inter-state crimes, the line that terrorism involves all aspects of government, and a bit of explanation about legal reforms.... Obviously, terrorism-related changes would have the topmost priority now. ]

—–

BTW, this is a minor aside, really speaking, but it would do well if our neighbours begin calling us “India” or “Bharat” rather than “Hindustan.” Pakistan’s rather articulate foreign minister was, just days back, shown on TV repeating “Hindustan.” This issue isn’t quite like the “Bombay or Mumbai” issue. It is deeper than that. India is neither a Hindu state nor does the idea of this nation state traces its roots to anything to do with those numerous “Stan”s existing in the world. The idea of India, the nation that we actually live in, today, just isn’t medieval… The expression “the idea of India,” which has become rather popular in the Indian press in the recent years, captures this spirit just right… It’s time we insisted on it… But, still, the first part of this post (regarding strengthening) is what is truly important; this small addendum, even if valid, isn’t, by comparison. 

—–

The Congress party has taken too long a time on both the counts: in getting Mr. Deshmukh to resign and then in deciding who the next CM will be. Bad! (As an aside, it also was surprising to see even “Sakal” Main Editorial attribute wrong words to Mr R R Patil… Why?) Poor show!

But then I just cannot attribute every bad thing to politicians alone (the way, for example, Sonali Bendre was found doing on TV this afternoon… I mean, does it all begin and end with just who we elect, Sonali?)… More on this tendency which the elites (intellectual or business elites) in India show, some other time… (I will have to go into many deeper issues like castes, socially en-forced—literally—artificial division of work and the related ethos, the resultant set of discordant values, the resultant deep distrust, the further complications due to mixed economy on top of this all, etc. etc. etc. Really, it will have to be a separate piece or two. Or hundreds… Some other time.)

The terror attacks on Bombay / Mumbai

This was the first time, I guess in all my life so far (at least since the COEP/IIT hostel days) that I watched TV for more than 5-6 hours / day. It kept my attention rivetted to the TV screen… It all was so shocking and plain different…

But war, was it? How can that be?… I mean, the question is: Can we give the respect due to a soldier—even an enemy soldier—to these terrorists?

The resignations of Mr. Shivraj Patil and Mr. R. R. Patil have brought out the essential difference of the Congress political culture from that of the rest of the parties in India. 

BTW, as powers that be (politicians, as well as the media [please note!] and govt. bureaucrats [please note this too]) come to occupy the center-stage, I now no longer watch TV. Before closing this para, I have to note a few things: (1) There cannot be any excuses, but I also found it disgusting that all the channels (without exception) put words in the mouth of RR Patil—things he did not quite say. In particular, he didn’t say anything along the lines of that movie dialog “Bade bade deshon mein aisi chhoti chhoti baaten toh hoti hi raheti hai…” What he did say, unfortunately (for him) is, by his exact words alone: ”Itne bade city-mein ekhaad baar aisa ho jaataa hai, iskaa matlab yeh nahin ki total intelligence failure ho gayaa hai…” But if you include *the way* he said it, i.e. his gestures, the emphasis of the unspoken words also into reporting—something that was not done by a single journo—it was clear that Mr. Patil was grappling to find words which would correctly express this idea that there inherently was something of an asymmetry in such kind of attacks… The poor fellow simply didn’t find the right words and didn’t have the politician’s presence of the mind—the thick skin, if you will—to handle this kind of a media situation… But at least, at the moral level, he made up for it through his resignation… What about the “worthies” currently occupying the opposition benches? Will Arun Shourie, for once, focus on them a bit honestly? At least for the moment? And how about media? Is it that their profession permits them to drop context anytime they feel like? Or is such a thing mandated by the lure of the lucre? (2) Here, the media also includes the goverment media (and bureaucrats also include the highest level ones at the central government level). For instance, Doordarshan news *showed* pictures with text of 162 dead, and then, 183 dead, even while the newscaster was, in his voice, actually uttering a figure of 195 dead… Why this fascination for 162 and 183?… Not just that news, but also a Secretary-level bureacrat was telling us 183 dead at a time that the JJ Hospital Dean was telling a figure near 120… I mean, do our bureaucrats and TV censors love to play not just with budgetory figures and the figures of estimates of monetary loss, but also with how many *people’s* lives have been lost? Has the cynicism (arising out of that irrational Indian fascination for, say, numerology or the lex-kind of “messages”) become that deep and ingrained? Mind you, the reports were honest for the first two days, when the action was still going on… The cynicism began to show its head as action came to end… (3) I do not further write about politics because it, and its expression in the media, are both too low… One will end up writing reams and still it would not be enough… (4) But yes, unlike last week, Barkha did conduct her yesterday’s show with dignity… (But she, and many journos, I think, labored unnecessarily long hours during the crisis… Don’t they have substitutes or shifts?) (5) Finally, a minor aside. I use the words “Bombay” and “Mumbai” interchangeably, and I mean the same city by either word. (I do not insist on any one usage but use it more or less randomly—on the basis of whatever comes at the tip of the tongue.) In India, there may be (nay, indeed are) many other cities that are both “modern” and “cosmopolitan” (as Prof. Dipankar Gupta today describes in ToI). But, Bombay is unique in that it has a specifically Western and capitalist character. It was founded in, grew, and acquired its definitive character in the ninteenth century. The men shaping it brought and imparted the specifically Western culture and character to this city. As the Marathi author P. L. Deshpande has pointed out, there was no Indian *city* existing in earlier times at the geographical location of Bombay. (This is unlike, say, Lucknow.) It’s entirely a British (read, a Western) creation. (… Much like Calcutta. Cal also is essentially a Western city in ways that Madras/Chennai was not and is not…. The difference is, Bombay has also always been far more Capitalist—it was not the capital of the British India, but a city that grew entirely on its own strength, i.e. on the strength of its world-wide trade.)… This all is so obvious, but then, what can an ordinary Indian intellectual do (and I do not single out Prof. Dipankar Gupta alone here) when he has a standing command, as if it were, not to identify anything Western inside India? 

—–

In some sort of irony, I happened to have written, right on Nov. 24, some thoughts on the need to strengthen the proper functions / institutions of the government in India. Such institutions included the police force. I was planning to post it one of these days, but then, this thing happenned in between… Guess I will now revise my writing before posting… The basic points will remain the same, but I will have to change the particulars a bit. I will post it here, may be a few days later.

Before I close, of course, my salute to those who died fighting this terrorist attack. May their souls, and also the souls of the innocent victims, rest in peace or experience “sadgati,” or so… My prayers go to them.