I just started a Chordoma blog with the first post titled, A Plea to the Chordoma Community. If you follow this blog because of my occasional Chordoma related posts, you will want to follow this new one instead.

I have spent the better part of my intellectually conscious life trying to algorithmically “beat the market.” (To be precise, I started around age seventeen; I am now twenty-five.) According to some very well-tested academic theories regarding markets, I am pursuing a fool’s-dream. I have continued to labor this long under the assumption that there is a pretty obvious selection problem when it comes to publishing findings that contradict financial orthodoxy: if you were to find a method that earned out-sized returns, I don’t believe academic prestige trumps monetary gains. It is my perception that people who end up as professors of finance are typically people who had the desire to study markets in order to profit from them, but who never found their holy grail. If they had found something spectacular, I don’t think the incentive to publish is very high. (There are exceptions, but nothing ground-breaking.)

After eight years, I have nothing concrete to show for my efforts. As a consequence of shifting needs, I have learned a lot of computer science (e.g. compiler design, algorithms, and some unnecessarily high number of languages). Obviously, this skill set is valuable, but I have no successful projects to use as credentials. Every few months, I find myself excited over the preliminary results of my increasingly sophisticated simulations, only to be disappointed a few short weeks later to find that I was simply wrong. This has happened so many times that I no longer grow excited when I see positive results — I’ve grown into a hardened, semi-depersonalized skeptic.

My latest iteration of development appears exceptionally promising, but I expect it to bear no fruits. I learn each time, and my understanding of markets (and complex systems in general) is approaching some level of refinement, but I have no way of estimating when I might cross the line into profitability; worse I may be approaching this level asymptotically, with my limitations acting as a ceiling just below my goal. I feel like a modern day Tantalus.

I recognize that “beating the market” algorithmically may be either impossible or simply out of my reach, but I soldier on because I still find it fascinating. I believe that, had I switched course years ago, writing off the project as foolish, I would have probably, or at least possibly, been wealthy by other means by now. (Every time I started pursuing a product development project, I found myself shifting back towards market soon after the initial new project euphoria had faded.) If I could offer advise to my younger self, prior to perusing this path, I’d probably say don’t make the attempt. I have neither ethical nor moral objections to profiting by speculation. I merely believe I could have acquired the satisfaction that comes from achievement a long time ago, instead of bearing the frustration that accompanies not achieving something in spite of my best efforts. Nonetheless, I will not stop trying. I’m not blind to the possibility that I am a smart fool, but I want this more than anything else. I’m not sure where I would draw the line, where I would finally say giving up is the proper thing to do. I hope I never have to make that decision. I hope success finally obliterates the need for that decision.

I’m not sure why I wrote this. To some degree, it might be a warning notice to those who are considering following this path. As I said, I find markets fascinating, but most people (my earlier self included) enter the fray believing it to be a sure and short path to riches; it’s not. There are far less risky paths to wealth, especially for entrepreneurial programmers.

Juno is a small Python framework similar to the Ruby Sinatra Framework. I needed a very simple way to look at a large set of png files in a specific way across a number of computers, so I opted to use Juno. It took me about twenty minutes to figure out how to send binary data dynamically. (That is, how to send image data via decorated function, not through Juno’s static directory convention.)

The following works, although it might not be the Juno developers preferred method.

def your_request_handler(web):
    content_type('image/png')
    append(string_of_binary_data)

It’s obviously simple but for me, it wasn’t obvious.

My first generation MacBook Air has been performing progressively worse over the past year. I had assumed it was just a consequence of software cruft, so I thought nothing of it. Then, last week, git commits started taking several minutes for small Django apps. It was time to fix the problem. Not wanting to be bothered, I did a simple OSX reinstall. (Although, since the Air has no optical drive, it was a bit bothersome; for some reason, the Air refused to communicate with my Desktop.)

Upon resinstall, I was quickly disappointed. The problem was not corrected. That annoying little beachball kept on spinning for practically every task. This suggested harddrive, so I took a look in the logs. S.M.A.R.T. was indicating imminent hard disk failure. (I really have no clue why no glaring alert was generated; I assume S.M.A.R.T. indicated the same thing a year ago — when I was under warranty — but I never checked the logs. Unfortunately, by reinstalling OSX, I lost access to historical logs.) After perusing a few articles online suggesting that replacing a first gen Air drive is very unpleasant, I just decided to bite the bullet and go to the Apple Store.

I went without an appointment, only because I was ignorant of their reservation policy. (Kudos to Apple on reservations; it’s a smart idea.) I was forced to wait around at the mall for an hour until a slot opened up. After learning the mall had no bookstore — a fact that made the snob in me feel angry — I turned to skimming HackerNews. (HackerNews is to geeks what Facebook is to non-geeks — somewhere to go when your bored and want stimulation.) Coincidentally, a blog post about hard drive replacements at Apple happened to be one of the front page stories for the day. This was fortunate, because it warned me that Apple would not allow me to keep my hard disk before having to deal with the genius bar “genius”; it spread my anger out over several quiet minutes instead of one loud minute. Thank you for making me appear like a decent person, HackerNews.

So basically, after deleting all my files and writing a quick python script to fill the hard drive with random numbers, I finally got to talk with my genius. (When you have a failing drive, filling it with random garbage is a painful operation to watch.) I must admit, Apple hires or trains employees very well. He was calm, funny, and generally charismatic. After I told him I want to keep my drive, he clearly explained Apple’s policy: they refurbish the drives and use “government standard encryption” to prevent data theft. By “government standard,” I assume they mean they overwrite the drive about a bajillion times past the point (one overwrite) where data could reasonably be expected to be recovered. I assume his answer is considered satisfactory for most users, but I was unhappy. I’m not paranoid enough to think my single-pass overwrite was going to make me vulnerable, but it’s still my drive. Furthermore, for drives that have not been overwritten, the eventual buyer is not the point of vulnerability — the many hands the drive passes through before being fixed are the dangerous ones. If the user requests their drive, not giving it to them is a bad policy.

Later on I realized that there is something more troubling about this policy. Apple, a consumer product company, is profiting on the failure of their own products. That’s pretty uncool. The replacement cost on my invoice was listed at $139.06 with an additional $85.00 in labor (HARDWARE REPAIR-LEVEL 1). By comparison, a new Toshiba MK8009GAH 80GB internal 1.8″ ATA-100 4200 rpm drive costs $114.85. Considering the cost of the very similar Toshiba drive, I don’t believe they were replacing my bad drive with a refurbished drive. (If they were, that would be especially ugly.) I assume they sell the refurbished drives as part of a refurbished MacBook Air, not as an individual component. Regardless, if the cost of a new drive plus shipping and less the profit from selling my drive refurbished is less than $139.06, Apple is earning money off product failure (i.e. shipping < refurbished profit). I’m angry.

Update:

A commenter on HackerNews asked why I didn’t just go to a non-apple service center. The truthful answer is because I am lazy. Nonetheless, it appears like the policy is set by apple and is universal amongst non-apple service centers. From MacService:

The old drive (working or not) is returned to Apple. If you wish to keep your original drive, Apple charges a significant core charge.

P.S. I’d be curious to see some hardware failure rate comparisons on Apple products. It’s my opinion that they produce bad hardware, but no one cares because most Apple consumers buy the latest version of whatever product Job’s slings out.

I am hoping to use Google AppEngine on a non-profit open-source project. To help me get familiar with AppEngine, I made a very small, simple webapp that sends a short message to any signed up users when they make a HTTP-GET request to their assigned URL. It’s useful if you have a very long running simulation or compilation that you would like to walk away from but be notified when it completes (or errs.) Assuming the email address used for your Google Account is also the one used on your smart phone, it basically mimics SMS without the headache.

The message should be shorter than 78 characters. Anything longer will be truncated. The message will be the email subject. To (feebly) prevent a quota DoS attack, each account is limited to 100 messages a day.

P.S. I learned only minutes ago that AppEngine apps cannot receive email from custom domain addresses, a feature I think my non-profit app will require. FAIL!

Learning a new programming language is fun. Different languages take different approaches to solving problems — they follow different paradigms. Learning a new language feels a lot like exploring, and I like exploring.

Exposure to more than one language is a good thing. However, once you are familiar with the broad paradigms, there exists a danger of conflating learning new languages with learning new ways to solve problems. In my case I forgot that programming languages are tools. Learning new languages can be seen as acquiring new tools. Learning a set of very similar languages is of limited utility because your not actually gaining anything new. There is almost no marginal benefit. To make things less abstract, compare the promisculous language learner with a carpenter. Carpenters don’t seem to suffer from the same error of judgement. There might be several brands of hammers offering slightly different features, but they all pound nails into wood.

I still enjoy learning new languages, but I don’t do so with such fervent desperation anymore. I’ve ceased my semi-conscious search for the best one. Instead, I finally realized my time is best spent studying algorithms, learning new programming techniques, and contributing to existing projects. This is much more rewarding than implementing the same software in twenty-three languages.

P.S. This post is a follow up to Consumerism and Programming Polyglots. I wrote Consumerism and Programming Polyglots a while ago. When I reread it today, I was very dissatisfied. It hinted at what I wanted to say, but it was inarticulate — mostly because I didn’t understand what mistake I was making by learning languages promiscuously. I think I get it now. Six month from now, I’m sure I’ll identify an entirely new aspect of my miseducation.

P.P.S. I wrote the title down, remembering someone had said it to me once when critiquing another post, Back to PHP. I searched my email before publishing this post and discovered it was Zed Shaw. This was disheartening. It took me several months since my original post to come to a conclusion that was accurately summed up by his flippant response: “All this fetishizing of technology is kind of pointless.”

The title is scraped from the link text I used to describe an article I wrote in 2004, during my first semester of college. I think I did it to satisfy a creative writing class assignment, only it wasn’t particularly creative since it was true. Reading it now, I take particular pleasure in the bits about “violently opposing” government intrusions into the internet. (I miss youthful indignation, self-righteousness, and the joys of masquerading as a rebel.)

It is also poetic in a semi-tragic way that, within a few weeks after writing this and resolving to “make a lot of money while doing something truly worthwhile,” I was diagnosed with cancer. It’s been nearly six years, but it looks like I might soon start making good on that promise.

The following is copied verbatim from the internet archives of that article.

Facebook redirects Konqueror to Facebook mobile; I assume it sees WebKit and thinks IPhone. Until facebook corrects the identification error, you can use Konqueror’s settings to fix the mistake.

Navigate to http://www.facebook.com/facebook?ref=pf.¹

Go to Tools > Change Browser Identification > Safari > and click the highest non-IPhone version you see. (On my computer, this was 3.2.)

Konqueror Facebook Screenshot

Enjoy lightweight browsing.

1. The setting we are changing will be applied to a domain. Since you cant get to facebook.com otherwise, you visit this page so that the modification is applied correctly.

Sometime last year, I had started maintaining email correspondence with a number of people from a diverse set of professions. Whenever I had an idea or a question that I was unable to fully flesh out or answer myself, I would find someone who was an “expert” on the topic and email them. Initially, I received replies, but they were terse. Courtesy dictated that they answer — especially those at academic institutions – but they felt no obligation to continue the conversation.¹

As time progressed, my messages became more finely crafted. My questions were more specific and the background material was assembled with more clarity. I learned to write better.² The act of writing diligently was in and of itself helpful, and writing with a critical reader in mind imposed diligence. I now offered value to the recipient, whereas before I was at best a nuisance. My ideas and alleged insights were at least well-formed, albeit not always novel. The recipients began to answer out of interest instead of obligation. The correspondence ceased to be one-sided — now, I had conversations and debates. At the point when I started BCC’ing people, I realized it would simply be easier to continue in blog format. Instead of updating a few people with a follow up email, they could just revisit the post and look for edits or comments; instead of emailing the same group of people every time I had a new idea, they could just add me to their Google Reader.

Like many nerds, I have attempted to create many blogs. (A small set, limited only by those I can recall instantly, includes: JustLikeJesse, ChasingSparks, jbn, and JohnBNelson.) Previously, I started blogging motivated by either AdSense or vanity — I wanted either money or attention or both. My interest in blogging waned shortly after the ubiquitous “Hello, World!” post and immediately before I found anything interesting to say. This blog, PathDependent, is the only blog of mine that has not floundered after one week — and it’s the only one where money and fame were never motivators.

Assuming I write reasonably well, my blog posts get attention. Attention by itself is useless at best while attention with comments is very valuable. There are limits to what I see in my own writing and thought processes. An idea may have such appeal to me that I completely ignore very relevant, seemingly tangential details — or major mistakes. Commenters — especially those that I do not know and thus have no requirement of politeness given anonymous commenting — have become unit testers for my ideas. Without commenters — private or public — I would not maintain this blog.

I blog because it helps me learn.

Notes:

1. Looking back, many of the emails resembled a parent handling a child who persisted in recursive “but why?” conversations. I’m almost embarrased by some of the messages.
2. HackerNews is probably more responsible for improving my writing than anything else. Conversation threads with karma acts like a unit test framework for ideas expressed in English.
3. I am defining success in terms of how it helps me explore ideas. The metric I use to judge success is the number of email messages elicited by each post.  I usually get a handful per-post now. While my traffic stats are beyond my expectations, traffic is valuable only insofar as it improves the odds of good feedback.

Edit: Josh Sommer directed my attention to How a Healthcare Company Can Accelerate Translation of Scientific Knowledge to Practice, which was mostly what I had in mind.

The mainstay of drug discovery is automated molecular screening, dose-escalating response curves, and clinical trials. It works, but it’s painfully slow and expensive. Meanwhile, patients with both chronic and acute diseases are continually given drugs with known efficacy. This has been going on for a very long time. Certain cost conscious elements of the health care system — health insurance companies — keep detailed records of this information on a per-person basis. I called up my prescription benefits provider last week and they provided me a copy of my historical prescriptions without a problem. I assume this is true of most, if not all, providers. I think these data could be very valuable.

For instance, I have (although there is no evidence yet of recurrence) a rare disease called Chordoma. If a sample of the Chordoma population’s drug histories were collected, certain inferences could potentially be drawn. Given known incidence rates for diseases, you have certain expectations. If you know 1:100 people taken should have had rheumatoid arthritis¹ but they are under-represented or missing, two things are possible²: people with Chordoma do not get rheumatoid arthritis, or people with rheumatoid arthritis take drugs that might incidentally treat Chordoma.

Does the sample of patients’ drug histories conform with expectations? If not, why? Considering the cost of clinical trials, this seems like it might be a low-hanging fruit, especially for patient-led non-profit research/advocacy organizations. I recognize that a straight-forward pre-existing drugs regiments are unlikely to be curative, but they may hint at promising avenues of exploration. (For example, does the group taking a drug for RA seem to have slower disease progression.) Such hints might not be possible to derive from in vitro and animal models.

Note: I am neither a medical researcher nor a medical doctor. I am just a guy whose reach exceeds his grasp. I took a cursory glance at PubMed and asked around to a few researcher friends of mine to see if they knew of anyone who tried what I suggested. So far, I found no evidence that this has been attempted. If someone wants to correct me, please leave a comment. If you know someone who might be able to correct me, please forward them this blog post. Thanks.

1. Adjusted for Age, Ethnicity, Sex, etc.
2. Well, several things, but two relevant to my argument.