Srikanth Chunduri visited me this weekend. The last time we met was in Hyd 3 years back. These days  Srikanth is up to living the life of a fast paced New Yorker in the financial district. He introduced me to some music from XLRI... enjoy.

If you don't understand Hindi, I would advice that you don't go about singing this without having checked the meaning of the lyrics with a friend. Enough said.

Here is another piece, which I would rather not embed on my blog:

http://www.youtube.com/watch?v=y_KT3sMR4CY

Enjoy.

After what should have been a brief hiatus that turned into a 4+ month vacation in India, Pooja returns to the land of McDonalds and "What's up?".

Pooja is not going to be Bloomington anymore - she is done with old IU - she is going to be at Seattle, working with the software company commonly known as Microsoft (again). With one small difference this time - Pooj is going to be a dev!

I just finished reading Mirza Muhammed Hadi Ruswa's 1905 Urdu novel "Umrao Jan Ada" as translated by the grand old man of Delhi, Kushwant Singh. What a terrible book!

Umrao Jan is based on the life of a courtesan in old Lucknow who went by the name of Umrao Jan. The book traces her life as narrated by her to Mirza Ruswa. In the process it gives you a glimpse into the life and times of the glory days of Lucknow and surrounding parts from the 1840s to the early 1900s.

The inside of the dust jacket says:

Old Lucknow is now but a dream, but this work preserves for posterity the full flavor of that golden age in the history of the city, which comes to life in the pages of this book.

We enter the palaces of wealthy nawabs, the luxurious abodes of cultured courtesans, and the hideouts of colourful vagabonds. The courtesans - out immediate concern - were on the whole well-read and exceedingly proficient in the arts of dancing and singing.

In the pages of this charming book, we find pieced together an exquisite civilization that will never recur, and warm human emotions that are eternal.

While this is all true, what is captured is not particularly beautiful. This is a book that seems to be a poor presentation of someone's (rather eventful) life. To its credit, I can say that it exists and so the story isnt lost. The book is littered with attempts at poetic verse that is, in comparison with some of the poetry that I have been hit by in recent days, just painful. (I guess you are set up for disappointment once you have developed a taste for the likes of Ghalib and Dylan.) The book also seemed to lack focus in some ways, unimportant characters held unnecessarily long conversations about unnecessary things. Lots of shallow moralizing.

In many ways Umrao Jan Ada has parallels with Geisha. Many similarities.

I havent seen the movie Umrao Jan, but I would expect it to be a far more pleasent experience than the book. A story about being "Lucknawi" truly has to be told in visuals with lots of flair and charm. There is not one, but two Umrao Jaan movies. The first one, the 1981 version, with Rekha as Umrao.

http://www.imdb.com/title/tt0083248/

The second, a glitzy 2006 rendition, with Aishwarya Rai as Umrao.

http://www.imdb.com/title/tt0485522/

I expect these to be better than the book - can't be too hard to manage that.
Enough said.

I walk into the Computer Science department today morning and I couldn't help notice little containers randomly placed on tables. I had a class on Computational Complexity and as usual I was running a little late so I decided to investigate later.

I walk into class, nod my usual apology to the lecturer and sit down. There on my table was one of them. Then I looked around and there was one of these on every table! Whats happening here?

They are little containers with unmarked white pills in them. Hmm... unmarked white pills? There is no text on the containers saying what the pills might be about. Strange. The only text on them says "JUST BE" in largish red letters on a white label under which there is a url in small font. Aaah.... the whole department was exposed to this stuff. And what a lot of victims they are, they are computer scientists - they hardly ever notice what they are eating. I would not have noticed myself had it not been for my temporary heightened sense of precense I acheive out of being late from time to time.

So what can this be? What sort of tablets would someone label "JUST BE"? A halucinogen? Some new form of crack? Birth control pills? After i went over all the usual possibilities, I started to panic. What if this is one of the secret government test thingies? Maybe a a neural inhibitor? or better a yet, a thought tranfer or a mind control pill? Then it struck me that this reminded me a lot of something I saw on the X-files many years back called "PURITY CONTROL"... JUST BE - PURITY CONTROL ... muuaaahhhaaa...

What sort of agency would expose the department? Maybe the whole campus to these? I decided to pocket one of these to examine it in detail later for clues. As I was doing so, I noticed the url - the url was for a cs department webpage for "WIC". I have heard that name before.. they did it! The "Women In Computing". The WIC exposed the department to unmarked white pills, which many of us would unsuspectingly eat. Why would the WIC want to do that? And what could be in those tablets? Come to think of it I dont know what the WIC really do... they are an organisation of some sort, yes (in most cases their membership is based on some marginal excuses). Why? Why?

This general general thought caused my mind to think about some WIC members. Aah.. well, lets skip that bit for a public blog shall we? And why has there been so much care taken to disguise these tablets to look like common mint? What would happen to me if I were to eat one? I decide to risk it all and try it. After all better take the risks here in the precense of people than to accidentally eat one one when I am alone at home with these tabs.

I flip the container over to try one and then I see this -

What? It expires today? Also it looks like it has been ripped open.. someone has eaten one of these! Something happened here because the victim is missing!

Then the sense and the safety in all of this becomes clear to me. It is a feminism thing.. and these tablet things are most likely just mint or mind control pills or something safe like that after all. And they were there as a show of good will. :) Some small details seem to have been missed, in this case the expiry date... details. After all who cares about details?

All this brought my Complexity lecture to an end. I guess I will have to find another muse for another day.

Khusrau rain suhaag ki, jo main jaagi pi ke sang,
Tan mora man piya ka, jo dono ek hi rang.

Thus, with the words of Amir Khusro, starts the 1978 opus of Shyam Benegal. I was so moved by this movie, all the little details in it, the characters, the poetry, the zeitgeist. The film aparently introduces the gorgeous Nafisa Ali. It also has Naseeruddin Shah, vibrant as ever, Shashi Kapoor, Shabana Azmi, Jennifer Kendall ...

The movie captures an angst in visuals and poetry, over the period of 1857 to 1858 - the days of the "Sepoy Mutiny" in a little town in Rohilkhand.

http://en.wikipedia.org/wiki/Junoon_(film)
http://www.imdb.com/title/tt0077783/

In a preachy mood: Today I spent some time talking to someone about how a piece of code that is going to affect his hapiness (aka he is responsible for) has some really bad design decisions in it. We went over how there is this little bug that shows as a consequence of this design flaw. The bug itself has an easy fix and such, but the underlying reason is that several layers of leaky abstractions interact with each other in unhappy ways.

In short I was trying to tell him that its not a bug, but that its bad design waiting for a victim - this should be better described as a pit or a trap than a bug. Bugs are visible warts, pits are lurking demons. Bugs are plain as day, easy to justify - pits are lurkers, those who can tell good code from bad, feel uneasy in their precense.

Do you have pits in your code? Do you have a sense of what parts of your code are pits? If you had to choose between fixing a pit and a bug, what would you do?

I believe everyone's code has pits to varying degrees and that they are completely unavoidable beacuse there is simply no such thing as a perfect apriori design. That said, here are some slightly more interesting questions:

If you had to implement something new - there is solution 1 that is going to introduce a pit but will deliver said target surely and quickly and there is approach 2 that will potentially eliminate the pit, but is going to be far more expensive to implement - How would you choose which approach to take?

I usually go ahead and implement the pit when I dont have enough information to decide between the two. Often there are pits that will never surface and no one gets trapped in them. When I dont know enough implementing the pit is a good way to (1) get my work done (2) ensure that I am collecting data so that the next time I know what the real costs are.

Its vastly more valuable to document the fact that a choice was made based on the lack of sufficient information, than the excrutiating details of choice itself. Pits turn into chronic illnesses of your code base when no one knows why ceratin decisions were made, and hence no one ever feels comfortable back tracking over a bad decision. Once I know an ill-informed choice was made, and once I know the price I am paying for it, its a lot easier to back track over it.

In other words, dont just document your design, document your choices in the context of your design space - it will keep you honest. Well written code to a large extent is its own documentation - what it is not a good documentation of is the choices that you did not take and the reasons why you did not take them.

Thoughts?