Apps
Contact
Seminars
RSS FeedFebruary 24th, 2010
A New Financial Model for YouTube
I posted a new financial model for YouTube in my software blog.
September 2nd, 2009
MP5 players?
Now a days, you can buy a cool 4 GB Video MP3/MP4 Player for less than 50$. Amazing price drop. However, some vendors are claiming to sell an MP5 player. Never mind the fact that there is no such format, and no such spec. 
When you read the details, it turns out that an MP5 player is simply an MP4 player that doesn’t require internal video conversion.
Just an example, how much obfuscation and confusion the marketing departments can create in their zeal to sell something.
July 9th, 2009
Google thinks Jennifer Lopez is a guy?
Now this is weird.
Try searching for “Jennifer Lopez presented her credentials to Steven Spielberg” on Google. It only takes the search king 0.17 seconds to shout back: ” Did you mean: Jennifer Lopez presented his credentials to Steven Spielberg ”
Well, no I meant exactly what I said!
I don’t really know what is Google’s issue, but somewhere between male chauvinism and idiocy is the answer. Of course, the information retrieval experts and other academics are shouting and firing up their Outlook to tell me about that the suggested terms have more search results. I see, I see. But I still don’t understand.
October 29th, 2008
Of geniuses and profound mistakes
Today, listening to the Frank Wilczek on the Kojo show, there was a one very interesting thing that Dr. Wilczek said, “Anyone can make an ordinary mistake, but a profound mistake requires a genius”.
That is an awesome quote!
By a profound mistake, he gives the example of Newton’s theory of gravitation, and how it proposed an idea of “instantaneous” force that is applied between two bodies. As we know very well now, Einstein’s general theory of relativity replaced Newton’s theory of gravity (read here if you want to learn more on that). I also agree with Dr. Wilczek when he says that Newton’s theory was basically “right for the time”; after all we still use it as it is much easier to calculate than general theory of relativity.
This idea of “profound” mistake is generalizable to other fields too – anything that solves a lot of problems but has its caveats is basically a “profound” mistake, and requires a genius. Such profound mistakes are sometimes the best things that happen to a field, since waiting for a “perfect” solution can take a very long wait. Examples of “profound” mistakes exist in other fields also. For example in the case of computer networks, TCP/IP may be a “profound” mistake. Of course, the protocols are not perfect, but they work so so well in so many of practical scenarios that they were truly one of the best things that happened to the field of networking.
October 7th, 2007
Get a life Apple fanatics
All the apple zealots go gaga over ithis ithat. When you read their blogs, they explain quite avuncularly that anyone who does not recognize in the beauty of apple does not have taste. This lack of reasoning is the true definition of fanaticism.
It seems that sometimes even seasoned bloggers lose the plot.
(i) For example, in this post, Joel Spolsky writes: “Ah, the way these five words (Designed by Apple in California) evoke a flurry of happy memories.” He proceeds to bash Microsoft’s line “Hello from Seattle”, likening it to Cobain. That is the best he has against Zune? I can see Joel filling out a music player survey, where on all the objective front he admits the iPod problems, but in the remarks section he writes “The iPod line evokes a flurry of happy memories.”
When Joel talks about happy apple memories, he certainly is not thinking about your cousin who got a shiny iMac, only to donate to the library 3 months later. (Except, that the library wouldn’t take any more macs.) He is also certainly not thinking about other iPod owners who can’t stand its strict police controls (or see my top 3 reasons why iPod sucks.)
(ii) Or as Carl Howe writes here: Why zune wont sell. He writes at the end “Full disclosure: the author owns Apple stock”. Since Carl has already confided that he stands to benefit if Zune fails, shouldn’t he also write in the “full disclosure” that: “Author has never seen or used a Zune”?
I am still waiting to meet an iMac/iPod owner who does not start his day with one hour of SeattleBashingYoga. There is a genocide in Sudan, a tropical storm over China (today), HIV is still growing, Iraq is still burning, and iPod owners are still going gaga over the fact that the iPod is sooo smooth that the fingers leave the prints the first time you use it. In turn, they prompt users who are OS neutral (like me) to respond to that.
August 21st, 2007
From not so good to not so great
Just finished reading “Good to Great” by Jim Collins, and I feel a bit underwhelmed.
My friend and mentor puts it a bit less mildly: “All business books start with a known result (success) and then work their way backward from that known point.”
I don’t completely disagree with that, though I think that books such as G2G do provide some insights into the business.
Let me just stop beating around the bush, and present a chapter by chapter summary:
- “First who, then what“: Good employees, will usually create positive value on their own. That being said, there are some good workers who may struggle to create positive value if totally left without any guidance. So, practically, if you don’t know “what”, it may be difficult to get the right people, and then engage them.
- “Consider the brutal facts“: This is the “backward” part that my friend talks about. The successful companies are called the “honest” fact facers. The unsuccessful companies are called arrogant.
- “Hedgehog concept“: Gist of the book. This is really important. Identifying golden circles of a company does provide good clarity to everyone.
- “Culture of Discipline“: Basically, this is just common sense that the discipline has to be in moderation.
- “Technology Accelerators“: As the author says himself, companies should use technologies to accelerate, not provide growth opportunities. Valid point, but a bit on the simple side.
- “Flywheel and the Doomloop“: Good chapter, a bit cheesy, but the overall concept is that the transition point may not well defined. Flywheel takes time and constant improvement, and that should be the goal, not a retreat that is going to provide a sudden lift-off.
- “G2G and B2L“: I don’t think anyone considers this a real chapter, its more like trivia.
- About 80 other pages at the end of the last chapter, such as FAQs, etc (not a regular book).
August 16th, 2007
Thank you for all the greeting cards!
Thanks to everyone for all the birthday cards – my neighbors, worshippers, friends, familymembers, mates, schoolmates, coworkers, classmates – for greeting cards, animated cards, video cards, everything. You truly made me feel special.
To top it all, it isn’t even my birthday.
(Seriously, what is up with all this SPAM?)
August 6th, 2007
Curse of Free Email
I will not start my tirade by wasting sentences about the curse of SPAM (how much and what kind of SPAM I get, etc).
Rather, let me reminisce. In olden (pre Internet) times, if you remembered someone (or you were forced by parents to), you would send a post card to your relatives. Or a letter. Or a card. Something. You would take a stamp, wet it with your dog’s nose and put it on the letter. Then, send the post card on its merry way. Would get there in a couple of days.
If you had *lots* of post cards to send, you may wish, wow, if it cost only 3 cents to send a post card, that would be much better. Say, your wish was granted, then you could send 4 times as many post cards. But you wouldn’t start sending it to people you didn’t know.
Today, with the new technology, the cost to send mail (email) has gone down to zero. Zilch. Nada. Therein lies the problem. If technology makes something cheaper, that is good. Making something totally worthless, not so much.
Say we had to pay 1c everytime we send an email. Looking at my outlook sent times and my OExpress sent items and my gmail sent items and doubling that all, I would have paid 22c yesterday! About 7$ total for a month. I would really love to do that, if in return the SPAM is eliminated. Now at this moment, a bunch of you are pressing Ctrl-N to send me a new SPAM email raising this question: (i) 7$ not so little for everyone, (ii) who would the money go to?
Ok, I have gotten your email, and here are my thoughts:
(i) If 7$/month too much money, how about 0.70$/month? That is 0.1 cent per email. Remember the problem is with spammers who send tens thousands of messages a day. It would start costing them 10$ a day. Maybe not a show stopper from them, but worse than 0$.
Also, giving some basic free (like some cell phone companies give first 600 minutes free).
(ii) SMTP/IMAP server. Money can only go to mail server host. They are the only ones who know when an email message is sent. ICANN could regulate that and collect some portion of the money from the mail server owners (and perhaps give a portion of that portion to some charity). Many of these mail servers are owned by companies such as MSFT, Yahoo, Google etc, so they would get some money too. Perhaps that would put an end to the term “free email”.
Wow, going back from free email to “very cheap” email. Anyone with me?
.
January 27th, 2007
Statistics Hacking – BBC as a Propaganda Tool
BBC is one of the leading news websites. It is often ranked in top 10 most visited websites. Due to this, it wields an enormous amount of influence.
BBC monitors the popularity of a news item on at least two counts (i) Number of times the news item was read, and (ii) Number of times the news item was emailed. These correspond to the “Most E-mailed” and “Most Read” segments of the BBC website.
As we show next, any hacker can propel a selected news item to the top of the “Most E-mailed” news story. The effect of this is quite significant and is self-feeding, as the news story that stays on the most emailed section then continues to get attention, and therefore continues to be emailed.
What is Statistics Hacking?
We define Statistics Hacking to be a process in which a malicious user manages to modify the system usage statistics. Statistics Hacking explicitly refers to the situation where the resource (website or system) is available to the malicious user for acceptable usage, but the user is able to modify the system usage statistics using some unacceptable methodology.
Basic Vulnerability in the BBC’s System
This section highlights in detail the vulnerability in the BBC’s “Email this to a friend” system, and how it can be exploited by Statistics Hackers.
As of now, BBC does not employ any of the advanced methods to prevent statistics hacking. Instead, it only uses a small hash value, that is hardcoded inside the HTML form. It is not clear if this hash was intended as a security mechanism at all, anyway it has zero impact in this respect.
Hacker’s Code
In the scenario below, we assume that the Statistical Hacker is a dedicated health services professional, who wants to highlight the availability of HIV Home Screening Kit, a news story carried by BBC http://news.bbc.co.uk/2/hi/health/6212467.stm.
Hacker begins by opening that page manually in a web browser, and then manually clicking on the “Email this to a friend” link. When the smaller window with email form opens, the hacker views the source of that page. The source of the page reveals most of the information that the hacker requires to submit that form.
Hacker’s Code involves a very basic Java program, in which a URL connection is obtained to the URL of the “Email this to a friend” page. Using the hidden variables and their values obtained from the form source, hacker creates the content that is then written to the output stream of the URLConnection object.
[Specific code for this story can be found in the PDF version cited below.]
Methods of Protection Against Statistics Hacking
Following categories of methods are available against Statistics Hacking.
- Use Computerized Turing Test to Distinguish Humans from Computers
- Use clustering techniques or improved counting to ignore double counts when analyzing statistics
Summary
This is a simplified version of an in-print journal paper. Full pdf version of paper can be found below.
Reference
[1] “Statistics Hacking – Exploiting Vulnerabilities in News Websites”, Amrinder Arora, International Journal of Computer Science and Network Security, March 2007.
Download PDF.