Usage is pretty simple, but has one caveat – to go to a link, you have to source the script file (from what I gather, that’s the only way to change the current directory by the means of a script. I hope this useful.

A very brief explanation of the code – it creates a file for every link you create, and the file contains the path. Note it might be interesting to add more functionality to that file, like a set of scripts to run on entering the directory, or something like that.

And yeah, thanks to Vimal both for Disqus and for inspiring me to put my ideas up online for open discussion.

Not All Those Who Wander Are Lost… But Some Are

“Have you heard about the murder case?”

“Hmm… It seems like I have a particularly nasty homicide on my hands.
Particularly nasty indeed…”

“What do you think about it, Father? You must know much about these things…”

“I have to admit, this one’s got me scratching my head pretty bad alright.
Sure, the modus operandi seems rather straight forward; it’s like so many of
those hit-and-run cases. But I dare say, what’s the motive? I haven’t a clue
who to suspect. Sure, I have his prints, but there are fifty thousand men,
women and children in London today with prints just like his. Or hers, I should
say. And a good thousand of those will leave tomorrow, replaced by ’round ’bout
the same from all over.

No, a person in my position needs far more to go on than prints. We need to put
our feet in that blasted man (or woman)’s shoes. That’s how we nab ‘em.”

“I’ve always wondered what goes on in the minds of such men. Murderers. That
needs some thicker skin, some colder heart don’t you think?”

“I’ve learnt about men I tell you. Homicides; they’re more about people then
you’d think. Sometimes I tell myself, “John, the reason you’ve done so fine
hasn’t much to do with your mind, son. It’s your heart.” People trust a
Reverend; far more than those ruffians in the Yard. Science and Religion, ’tis
a heady mix if there ever was one. You’ll have the heart and minds of people at
the same time; trust me it’s worth a thousand times more than either alone.”

“Hmm… A thicker skin. I don’t have one of those. Don’t you suppose one could
perhaps, kill a man out of… oh, just entirely by mistake? Haven’t there been
cases when a jolly ‘ol matte smacks his drinking partner hard on the back and
that causes the pour soul to chock and… and die? Is that ‘ol mate a criminal?
Who knows what circumstances cause other men to kill?”

“Criminals, they come in all forms; different shapes, sizes, motives. Most
cases I hear are just out of anger, jealousy, some such mundane emotion. The
cases that really terrify you are ones caused by some innate wild
uncontrollable animalism. They terrify you because you know they just won’t
stop. Not till they’re dead. The others you can reform, hell most of the time
life in jail is better than what they have out here. Sometimes, well, most of
the time though, they’re easier to catch; the perpetrators are far too sloppy,
and frankly, stupid.

There’s a final category of people, which I’m yet to come across. The
aristocratic psychopath. Cunning, intelligent, resourceful. He’ll do whatever
he wants to, and can throw his money around to get away with it. He’s smart
enough to make the case look like a complete accident; like this hit and run
case. You just never know whether what your looking at is a masquerade for some
greater scheme…”

“Does everything happen for a reason? Do all men have purpose? Couldn’t one
man’s purpose, as it were, be to end another’s life? How would such a man feel?
That his Purpose was to end an man’s innings upon this world? I’d be pretty
angry to have to sign that contract with God. Surely something must happen just
by chance; devoid of reason, and hence of guilt?”

“Nothing happens by coincidence. Not for me. Why, now, would Britain’s most
wanted, an adept criminal by all standards, one who’s been a greased pig for
every detective in the Queen’s country for over a year and a half now, be lying
drowned in a cask of some rather crass lager? In a black as night alley along
the River Thames? Under a broken street light? No exciting last chase as the
Yard slowly begins to corner him, no famous last words, no stunning
revelations. No dignity. One of the greatest criminals of this decade, and I
dear say this century, dead without any last tricks that fail by just a
whisker. And you expect me to believe this is not the art of some even greater
mastermind?”

“Tell me Father, is it a crime to thieve from a thief? Oh, I know of ‘honour
among thieves’, but if I were not a thief? Well, I suppose it may depend on
what was stolen. And if it were something of as great a value as one’s lives?
Several lives even?”

“No, I think not!”

“Then would one be justified in the murder of such an individual? Surely the
police wouldn’t care much to find the murderer of a murder? Shouldn’t he be a
hero? Hmm… People always love a vigilante; though what the authorities think
is a whole new matter…”

“As a detective, it’s my sworn duty to pursue the truth of the matter. Am I
glad the devil bastard is dead? Certainly; but to not follow up the case is yet
a greater folly. Who knows what the killer of the killer may be up to next. It
is more likely than no that this Mastermind is simply eliminating his pawns.
Hiding his pieces so we may never see the player behind them.”

“I’m no criminal. I haven’t the thick skin, see, mine is as soft and supple as
that of a scholar’s laddie. And see my heart? As warm as the midday sun. I
couldn’t be capable of Murder of all things.”

“Now, I can’t go about suspecting anyone. This is a tricky game, yes it is -
twists your mind it does. If this Mastermind is mastermind-ly enough, the
suspect won’t even know what he was doing. He could make it seem like a pure
coincidence. I dare say _you_ could be the criminal, and not know it. Would
that truly make you a criminal? I don’t know”

“Just imagine, committing a crime you did not intend to commit. Imagin the
thought of having another man’s blood on your hands. How would you feel? Should
your life stand still, stop in it’s tracks? And it wasn’t even your concious
doing…”

“No, I suppose not.”

“But still, the unholiness of it. Wouldn’t you feel as if God had taken smite
on you? Wouldn’t you question yourself, put yourself through anguish, try to
find what sequence of mistakes had caused His retribution in such a form.
Surely, He would not condemn a man to murder were that man not condemned
already?”

“This man, the Mastermind, he must be a truly remorseless chap. I may err in my
reasoning, but no women could be quite so heartless, to kill pawns with pawns,
so impersonally.”

“Does the Lord have any mercy? This man here, has no great fortune, no wife or
children even. And you take away his only hold on reality, his values and faith
in you?!”

“An aristocrat. A scholar. One of the two he must be. Oh, manipulation of men
and their thoughts is inbred in them. I am certain of it. What could such men
want from the world? Wealth and power they have already. Women, well, women
can’t be got this way…”

“Goddammit, I’m just a boy. Driving my cart, doing my father’s chores. I
didn’t indent to roll off the curb and hit that man. Get my cart jammed in
there like that. I tried getting him free. God didn’t even bless me with the
strength to save him. It’s not my fault!”

“I should start by checking reports of all the lord’s in this area. This is not
going to be easy, no. Wish me luck, ‘ol chaps, I’ll need all there is to solve
this case”

“Sigh… Forgive me Father for I have sinned…”

So a couple of points that I really like about writing the story:

• It was entirely in dialogues, that too between two characters in completely
  separated physical locations. That meant I could be inconsistent, and didn’t
  need to focus on my language, as long as I kept a consistent tone. And it was
  fun sketching out characters entirely by their words, and trying to make their
  dialogues roughly match each other.
• This was one of the few times I’ve had to write a story, by hand. I generally think of the story either non-linearly or hierarchically, thus writing out bits and pieces and joining them together. I blame lots of the inconsistencies in the story on it (of course, it couldn’t be my fault otherwise).
• I threw in 5-6 themes of the “suggested” list (things like the title, the “setting”, etc.)
• I had initially planned to make it something along the lines of the dialogues
  in Crime and Punishment (Fydor Dosteovsky), where the criminal describes
  exactly what he’d done to the detective, but because it was a murder of chance,
  and rather unconventional, for most part of the story the detective is utterly
  clueless. Somewhere along the way, I decided to separate the two physically.
  Atleast initially I had planned for the to-fro dialogues to be roughly
  coherent, but for some reason, I decided towards the end to skew them apart as
  much as possible to highlight the irony of the situation, and also to be able
  to raise to pace to a climax.
• To be honest, it sounded a lot more coherent and “cool” when I was writing it. I swear.

Edit: Corrected some spelling errors pointed out by Suresh Govindarajan. Also added a point to the commentary

I would also like to shout-out extreme gratitude to the designers of the Simplr theme, which, with a few minor modifications, fits with my main page theme quite well. Woo. Also, a shout out to favicon.cc for a CC-licensed favicon that I modified on the site itself.

Oh yeah, and about the content, I’ve been dry as a desert lately because of a) lots of work, b) a lack of need to express myself, and c) insufficient E_a. However, I have a couple of random thoughts and introspections, apart from technical experiences that I expect to jot down here soon. And research-ish ideas (which shall see public discussion thanks to Vimal).

The overwhelming focus of this year’s foss.in was open hardware – something I’m really close to. The best talk I attended without a doubt goes to Harald Welte; he spoke on how people (including himself) had reverse engineered RFIDs, and remarkably GSM, and exposed how flawed it was. Harald has this near holy modesty about him, but good god does he know his stuff. Seriously, Harald ist Gott.

Now, the other set of great talks were given by the anti-thesis of modesty, Lennart Poettering (who ist also Gott) (who I’ve venture to call RKK’s protege). He ripped apart any “assumptions” you could possibly have about filesystems, S.M.A.R.T., operating systems, hell anything. General theme (as always) was “Everything sucks, and here’s why.”

Coming to hacking, I must say, GNOME day was a bit of a disappointment. The “focus” this year was GNOME performance again. Some interesting work was done using dtrace on Solaris (now that I’ve seen the tool, I have to admit it is awesome). I sadly didn’t see much cohesion, nor any direction with the GNOME community, which kinda irked me. Personally, I spent the better part of 3 days profiling, checking out applications that needed performance updates, etc. I did find a performance bottleneck with GTK, but it had already been fixed (of course it took me the better part of a day to get the latest version of the sources to verify this…). I also gave a ad-hoc talk about jhbuild and getting “bleeding edge GNOME”, which I think I utterly mucked up, and would like to forget about. Only patch I submitted was one a trivial one to jhbuild. In other words, foss.in/2009 was epic fail for me.

On the other hand, the KDE community was awesome – they had a tight show running, and managed to guide quite a few people into development, and I’m pretty sure a few of which will stay. I bow to Pradeepto, and Akarsh.

On a concluding note, I’m currently having quite mixed feelings with my role in FOSS – I’m certainly not “contributing” anything significant, and I don’t feel much of that community goodness. I think I’m going to probably stick with just opening all my work. It seems rather ironic that my first foss.in had pulled me into FOSS, and now my second (perhaps last?) foss.in marks the end of that period.

The concept of a generator may be very foreign to those who haven’t really explored Python – though it is a general principle (like iterators). The crux of it is to generate the members of a set/relation one at time (which are subsequently returned). If you are not going to store the enumeration, this methods prevents needless use of memory (that may lead the program to be extremely slow). The trade of is the (minute) expense of storing some “state” in the function, and perhaps some additional function call overhead that is rarely a concern.

An Example – Lexicographic generation of strings

Let us begin by taking this example that is used to enumerate over all strings of length n in lexicographic order:

Now for some back of the envelope calculations: If you ran this code for strings of length 5, you would be generating 26^5 = 11881376 strings each of length 6 (because of the null char) which amounts to 67Mb. One character more and you’d need 1.8Gb. Considering that you are going to throw this away immediately, clearly this is not the right approach. The “usual” approach to solving this problem is to include your code in the middle of this loop like so:

This is clunky, ugly and non-generic, in other words evil. Another method you might devise is to use a “static” variable – this is of course for those of you who think in C.

Python introduces a beautifully clean solution (as usual) – one word: yield. The yield keyword basically saves the state of this function, and returns to you the computed value. The first time you run the function you get a “generator object”, on which you call the “next()” function to step through (the generator object is also iterable by the way). This modification allows you to compute any length strings in a breeze (much thanks to Kirtika for catching that my original post had “list += str” and “return list” in it instead of the yield command. Talk about a screw-up…):

For comparison, this code was used to generate SQL files for an assignment in our databases course. The earlier version took greater than 10 min. to run (before exhausting all my memory mind you). The newer version took about 10 seconds.

Enumeration of Infinite Sets

A key result of the generator model is that it enables you to enumerate/work on potentially infinite sets! For example take this example to enumerate all even numbers (I agree this is a pointless example, but bear with me):

This is something you just can’t do otherwise.

Generators in C++

For the sake for completeness, I’d like to touch upon how you could implement such a model in C++. Being object-oriented, C++ also lends an elegant, though not as succinct an implementation – Functors (aka function objects). This obscure functionality of C++ is terribly powerful, but I’m quite certain that few C++ coders have ever heard or considered it.

Notice the semantics of the operator() function. If you understand operator overloading, it’s pretty clear what’s happening here.

Now, as I mentioned earlier, the core behind a generator is storing the intermediate state of a computation. Python handles this internally. If you wish to do the same in C++, simply define a function object that modifies it’s private variables every time operator() is called, like so:

Every time you instantiate an EvenGenerator, you’ll have a shiny new object with an initial state, and each object is independent.

Generators in C

C does not lend itself naturally to such a paradigm, and any implementation is intrinsically clunky (as usual). While static functions “work”, they do not lead to re-entrant code. Before I move on to what I’ve thought of, an example of a very clunky implementation a “generator” in readline’s autocomplete example (perhaps more details in a future article):

This is still not re-entrant as only one person can use the generator at any given time. A possible implementation to solve this is to store the intermediate state of a computation in a struct, and pass this to the generator function each time:

Sure it’s clunky, but that’s C for you.

This list spans an infinity, and having much to say about each I’ve decided to start an “Excursions in Programming” series to cover them in sufficient length to do them justice. While these techniques may be nothing new to those who’ve used it, I think they serve as eye-openers to those who haven’t. I hope that this collection will interest those few readers who share an interest in programming.

What’s in Store

1. Generators in Python, C and C++
2. Lambda functions in Python
3. The almighty __setattr__ in Python
4. PDB and Python
5. Private functions in Python
6. Using readline
7. Notes on Yacc/Bison
8. The Power of Git
9. Valgrind, GDB GProf

I’ll probably take my time to generate this whole list, but rest assured that I will have less technical topics in between.

A nice question that Ramya asked was not the standard “Could computer programs even be capable of creativity?” (which is IMHO a completely philosophical discussion and I side an author a work I read – it entirely depends on how you define creativity), but rather, “Why would you want ‘creative’ programs in the first place?”. Computer programs that detect or comprehend things that humans can do (like how programs can detect handwriting, or pictures, etc.) are immensely useful to us, it helps simplify a lot of routine work.

If you look at the point of human existence as an elaborate ploy to create things (and subsequently destroy them), as I do, “creative programs” seem like a step in the opposite direction. Does “creativity” have any meaning to us if it doesn’t involve our own ingenuity? Or in other words, do we appreciate “creations” because of the mystery in their happening (as we do with nature or other artists?). I think not – what difference should it make whether creative arts produced by a “computer” or anyone else. Of course, computational creativity is hardly good enough to actually be appreciated (except in art, where there are some really good results (“Aaron the Painter” and Wikipedia).

My interest in the topic is purely academic, and the preceding discussion is more or less for amusement. To me computational creativity has 4 raisons d’etre:

1. To better understand the process of creativity.
2. In all likelihood we will find that computational methods can not produce human-like output (perhaps it is foolish to think they should be human-like – perhaps it’s output is “better” – though I doubt it), and we rejoice our uniqueness and awesomeness and go own with our inane item songs and pointless reality shows and whatnot.
3. To exercise our own creativity in creating something capable of “creativity”
4. (My favourite) To take our own creativity to the next level. What new majestic works of art would be we be capable of if we had help in our planning and execution? Could we plan or design a grand musical tapestry at a high level and fine tune computationally “created” output generated?

A concluding note, it often seems eerie how we categorically dispel any sense of “self-importance”. We are not at the centre of the universe, we are nothing but evolved monkeys (amoeba even), we are largely made from the same chemicals that make up lipstick and Garry Kasparov could be defeated by a computer program…

The projects that were being hacked on this year were:

Rumours had it that KDE was also present at the hackfest, led by Akarsh Simha (kstar), but I think they never managed to get KDE to build .

The FOSS events at Shaastra this time started off with Atul Chitnis (foss.in) giving an introductory talk about what FOSS was and why one should contribute, shortly followed by Shreyas Srinvasan (shreyas) giving his most awesome “FOSS Foundry” talk (the same as last year, only this time we had a net connection going, and things were a lot more streamlined than last year). Later in the evening, we had a brief introductory session where each project “leader” spoke about their project.

During Night 1, there were some initial hiccups in the setup, and build systems, but they were quickly resolved by the truly amazing Subhashini (our super-’vol’), and covered up by a rather arbit talk about IRC by me. The groups were split up, each group occupying a different part of the department. Unfortunately, since I was organising one of the hackfests, I didn’t really get to check out the awesome things the other groups were doing. I do know that sid0 was working on making Firefox plugins using Jetpack, Jai introduced participants to ffmpeg internals, and Akarsh was basically compiling KDE.

I started my session off as a GTK+ workshop, working my way up through getting help for yourself, what the point of function pointers was, makefiles and finally how the whole GTK main loop, etc. works. It was all ad-hoc, and followed a path best modelled as a Brownian motion. However, the objective of the first night was to equip the participants with the tools they would need to work on GNOME code. It was really heartening to see participants try interesting and ambitious stuff with simple GTK+ applications I had asked them to make. If nothing else, I feel confident that atleast 10 more people have become GTK+-literate, and I think that’s a starting point for a lot of awesomeness.

Night 2’s goal was to actually get participants to hack on simple bugs. The plan was to work on removing some deprecated symbols, etc. as listed on GnomeGoals². I did give a demo on what it takes to create a patch for GHex and this GnomeGoal. In one of those amazing hackfest moments, one of the participants found a bug in GNibbles – the scorecard/preferences, etc. dialgos are not modal. This meant that you could change the game’s settings while the game is in progress, but of greater alarm was the fact that the game’s state machine to pause / unpause the game would get f00ed if you opened the dialog boxes before starting any game. This program was easily remedied by making the dialogs modal. It was but a 3-4 line patch, but initiated two rather awesome chaps – Bhargav Prasanna, and YuviPanda – into GNOME (Bug #531825). Bhargav Prasanna continued to work on another bug, this time entirely on his own, this one was: Bug #323309.

I was also a participant in the Open Embeedded workshop/hackfest that happened on Day 3. I’m sure Kirtika Ruchandani will post more about it, but as a participant this time, I can say that it was tremendous fun. Most of the bugs were more trivial, and less involved than those found in raw code, and that meant that we could fix them without know a thing about OE . Subhashini (the super-vol) and I worked on a couple of bugs reported to the tracker, and she got to submit her first patch (Bug #5325 – it hasn’t been committed yet). We also spent some time coming up with a bash completion script for bitbake (#5326 – also accepted, the OE community is fast).

A brief shoutout to the amazing people that made it happen:

• Vinay Hegde, Prakash Mohan, and the rest of the Hackfest Team: They organised the whole show, putting endless nightouts to get things working.
• Akarsh Simha (kstar): He was responsible for getting the whole team fired up, and moving, and putting endless nightouts to get KDE to compile.
• Subhashini: The tireless super-vol was responsible for setting up the whole lab, and made the hackfest happen. I can safely say that without her dedication and expertise, we would have been a struggling no-show, some what like how last year had turned out. Infinite props to her.
• Akarsh Simha, Siddarth Agarwal, Jai Menon, iwikiwi, Kiritka, and the IBM guys: They flew down here to lead hackfests for their respective organisations, and really added zing to the show

Here’s a list of other peoples’ views on the hackfest:

• Akarsh also has his review up here and here.
• YuviPanda has this and yet another.
• Bhargav Prasanna’s seems to have made it to LFY!
• Jai Menon’s

(Much thank to Kirtika for coming up with this list)

Footnotes

¹ Contenders to the statement are more than welcome.
² A brief aside here – GNOME’s wiki brilliantly documents simple tasks that need to be done. I am particularly impressed by this link which dynamically populates a list of GNOME projects with how much deprecated code they use. It was tremendously easy for me to find out what tasks we could do in our hack night.

The "fearsome" Dutch Cemetery

Having decided to go on the trip only about 8 hours before actually leaving, we didn’t do much of that “planning” business. We just bought some grub, and took a map of the route there (courtesy Google Maps of course), and got out Coco’s trusty (yet dusty) ‘ol video camera. Our famed “BBz” was sadly reduced to the less prolific BBz-2 (for the record, one of the BBz ditched us, and the other chickened out. For those who know who we are, which one did which is rather obvious). We set out at 4:50 (am – yes we can get up that early), only 50 minutes from when we actually planned, and spent a good hour cycling in the wrong direction (my fault entirely). Post some course-correction, and about 2 hours, we had explored all of the Chennai coast and had vented our anger at it’s utter lack of drainage systems. After another 3 hours on the highway, we rode into town.

Few words can describe Pulicat. Not to throw you off, this is simply because there is very little to be described about. It has neither a lodge, nor a restaurant. But since we were riding Tam cycles, and spoke “Tam”, we blended right in, and made ourselves a deal for a place to take a bath (thank god for that), some grub, a boat ride across the lake, and a tent to crash on the beach with. Of course, being “Tam” doesn’t mean you don’t get scammed.

Probably the only “thing” to see there is the Dutch Cemetery, which is singularly boring as well. We began our video diary there (so as to scar future generations), took a couple of videos along the way, and for the rest of the trip. I got to steer the boat for a short while (ooh – caught on tape), and we setup tent on the beach. Let it be known that there is nothing more awesome than crashing on the beach, what with that blissful sound of beating waves, fresh breeze, and constant fear of being sucked in by the tide. I don’t think I’ve ever crashed that well before – it made the whole trip worth it.

Next morning we set out at 6, and made it back to the city by 9 or earlier (that’s 45km folks). Unfortunately, we got stuck in the city and it’s horrid traffic – swore in public (loudly), experienced road rage on a whole new dimension, got harassed by a cop, and finally made it back to insti. In summary, we put infinite fight to cycle 60 km to the middle of nowhere, crash there, and come back. There are somethings that you just can’t change; we just had to maintain the time-honoured principle of the BBz – Veni, Vidi, Dormi.