So with the advent of Google Code Search, I’ve tried at first to search for my name. Sadly, nothing comes up despite my having contributed to sourceforge projects, and having started one of my own. So I did the next best thing, I searched for the more interesting ones. Here’s what I learned:

• bar (3M) beats foo (1.5M)
• There are hacks (600K) and then there are huge hacks (100)
• There’s lots of broken (542K) but there’s definetely not enough bbq (6K).

What other interesting searches have you raked? Post a comment!

There are certain reasons why I admire Google more than any other tech company out there, and one example is their Summer of Code program.

Under this program, a $4500 stipend is given out to college students who pick one of many available existing open source projects to work on for the upcoming summer. The projects, which have usually no established sources of income nor any personel on payroll, are as important as Subversion, Firefox and KDE, though the list is very extensive and even includes projects you’d like to do yourself, under Google’s own mentoring.Not only does the program support much needed open source projects by providing talent, but for every student mentored, Google will also pay the mentoring project $500.

Needless to say, if you are a Computer Science college student and use or write open source software for fun, this is an amazing opportunity. The pay is quite good for a job you can complete at your own pace, and of course since most open source projects have no offices to begin with, you never have to show up at work! And should you do a good job, you just might spark some interest in yourself from the folks at Mountain View; the very least, you get a T-shirt, a certificate, and the satisfaction of having worked on some of the best software and with some of the smartest people in the world.

Apple shares declined a whopping 6.33% today, possibly on rumors that sales for new macs might be hampered by a lag from software developers to switch to the Intel (INTC) platform, as well as news that CBS will be bypassing iTunes and will provide direct downloads of the show Survivor.
I have been watching the decline on intraday trading today trying to figure out why the sudden selloff - but I could find no major coverage on either one of the above possible explanations on the regular sources, among them the Dow Jones newswire, which only covered the CBS story.

The Survivor news might reflect a reasonable fear that the iTunes dominance might be short lived; I find that unlikely:

1. Survivor is the only show with which CBS is experimenting with direct downloads; it looks like its iTunes for everything else.
2. iTunes’ revenue is still primarily music, to the best of my knowledge.
3. People are used to iTunes; it’s easy, highly available and accessible, and works well with the iPod.
As far as the Intel switch goes, it looks like the primary concern was Adobe. This is nonsense; Adobe must have been waiting for years for this switch. Its products sell well for Macs, but they sell the very same identical products for Windows. Eliminating the PPC support will allow Adobe and all the companies that sell for both platforms to consolidate their codebases. This is easier than you might think - worst case, they can run their software under an emulated mode, and speed concerns might not be an issue since the Intel processors are, if you were to believe the hype, 4 times faster than their PPC `equivalents’.Let’s consider the other end on the other hand. The move to Intel allows a myriad software providers to publish on the Mac. Should things work as planned, we could see PC games being ported to the Mac, allowing Apple to tap a whole new market that has been ignored for years.

During my Fall 2005 Semester at Hopkins, I took Database Systems with professor David Yarowsky. The course included a final project, which I did with Raymond Buse. Our aim: to build a simulator capable of investigating the performance of technical analysis tools over a long period of time.

Technical Analysis is the study of the price movement for a security in order to extrapolate a dependable trend that could theoretically predict a future price and allow the investor, or as Benjamin Graham would correct, speculator, to act upon this prediction. It is sometimes referred to as chart reading or voodoo finance and contrasts with fundamental analysis, which is the in-depth study of the security, such as the profitability of the underlying company in the case of an equity.

Needless to say, this practice is often the anathema of seasoned investors. To understand why we decided to quantify how a trader would perform if the decision making process relied exclusively on the output of these algorithms. As control data we had two non technical strategies, BuyAndHold and InsiderTrading. The first is self explanatory; the second relies on an oracle of perfect information that knows the exact stock movement of the following day for a particular stock, and performs a transaction only if it will be profitable the next day.

Our results showed that starting with only $10,000, you could buy Microsoft in less than 10 years regardless of when you started to invest using perfect information, however, there’s no stock on the Dow Jones when you would be able to consistently beat the market in the long run using any technical analysis and that in fact, BuyAndHold would beat the other strategies. The results for this study will appear sometime soon on this website.

Given how versatile this simulator is, another interesting study would be to compare Dollar Cost Averaging versus Lump Sum investments; perhaps this will also be done soon, and if so, the results will be posted on this site.