MarkPNeyer.com

As it turns out, a for cylinder to be ‘capped’ doesn’t mean that it has a flat top, like every other cylinder I’ve ever used, but rather it means it has a rounded top. Like a sphere that’s taken in the middle and stretched out. If you wondered, capped cylinders make lousy tires.

Tonight I started work on integrating pyODE into my game. At first I thought it was going to be very difficult, but then I started working on it and it seemed like it was easier than I had expected. Then I tried more and found out that it was, indeed, rather difficult. I will prevail, though.

There’s something I enjoy about finding bugs in a program. The methodical search for the imperfect delights me, and whenever I find a bug my mind usually works on several possibilities at once. As a general rule of thumb, when I see my program behaving strangely, in two ways that seem totally unconnected, it’s a good posibility that they are linked directly. There’s something very intellectually pleasing about two unconnected phenomena that are actually directly related.

On that note, I beleive that computer science and programming are the purest of all sciences, and that programmers make the best scientists. The reason for this is simple; science is all about the scientific method. First, some sort of phenomenon is observed. In an attempt to explain the phenomenon, a model is constructed or an explanation given for the phenomenon. In order for the model to prove usefull at all, it has to be tested. If the model withstands testing and results are produced succesfully from that model, it’s accepted as ‘true’ or ‘valid’ or whatever you please. If it doesn’t produce usefull or valid results, it’s rejected and a new idea is sought out.

Finding bugs in a software program is science on a small scale. You observe a bug. You construct a mental model of your program in order to determine what caused it to err. Once you have constructed a model and explained why the software is doing what it’s doing, you test your hypothesis by looking for and fixing the fault. If you run the program and it works, good. If not, you revisit your mental model and attempt to correct it in order to have it predict the bug that you’re observing.

Another thing that I should think would make good programmers into good scientists is the nature of the debbugging proces – I’m not sure how others work, but for me the key to understanding why a bug is happening is to insert all kinds of output statements in my program, so it tells me exactly what it’s doing. This is akin to constructing an experiment and gathering all of the necessary data. With a computer the only data you can’t collect is that which you don’t understand.

On a final note, the biggest difference between computer science and the rest of scientific endeavors is how easy it is to ‘jump in and get going’. If someone tells you that an object is made of up atoms, you can’t bust it up and count the atoms. You can’t take it apart and put in different atoms to see how it works. You can’t even look real close and see the atoms – all you can do is take his word for it, and beleive him when he shows you pictures of some little bumps. With computer science, that isn’t the case at all. If someone tells you that this sorting algorithm runs in this time, you can easily implement the algorithm yourself and see how fast it runs. If someone tells you that acecss times for heaps are on the order of log(n), you can check it out for yourself. True, there are some things like parrellel computing and such that you can’t really do unless you have the equipment, but for the most part all you need to do computer science is a pencil and paper. Sure, a computer helps, but it’s not really even necessary. Most of the time you can even do it in your head.

I could go on, but I’m going to get back to programming.

The title applies if you’re interested in this sort of thing, only. If not then maybe you’d like this.

Consider several .mp3 music files stored in a .zip file. What’s going on there? What’d it take to get those files there? First, some source had to produce rapid fluctuations in local air pressure, with frequencies ranging from 50 to 20,000 times a second. These fluctuations in air pressure caused a distrubance in an electrical system, most likely moving some sort of magnet through a loop of coil, causing current to flow opposing the change in magnetic flux. This change in current passed through a single cable which was attached to a digital volmeter. At some frequency, most likely 44,100 times a second, the output value from this voltmeter was shunted along a data bus to a memory controller coupled with a simple counting circuit which increments is incremented at that same 44,100 frequency. Every time the voltmeter’s output is latched to the databus headed for a memory counter, the counting circuit is incremented and the current voltage on the input signal is stored in the memory in binary form. While this process is occuring, a microprocessor somehow connected to the aforementioned memory controller ocaisionally follows an instruction whereby the entire contents of that memory controller is latched to a databus, fed into a register somewhere near the procesor and then moved into a hard drive’s cache for permanent storage. This process was repeated several times, to create several raw audio data files, which are a record of measurements taken of those air-pressure variations. By creating air pressure variations of the same intensity recorded in those data files, the sound of that disturbance could be reproduced.

Once those audio data files were created, a new aglorithm is run upon them, sliding a short ‘time window’ along them and using a form of Fourier transformation called a ‘Modified Discrete Cosine Transformation’ in order to roughly approximate the signal in that short window with a series of sine and cosine waves of varying intensity and frequency. The result of the aglorithm is information describing a mathematical function which, although not identical, gives a good local approximation of the original signal made by that source above. This new information is then scanned by a tokenizer which makes frequency counts of tokens of varying lengths found within the data file storing information that describes the rough approximation of the original squence. The most frequently occuring tokens are stored in a list and instances of these tokens are replaced by an index into that list, resulting in an even smaller amount of data. This data, when certain token references are replaced by their values from a lookup table, can be used to construct a mathematical function that closely approximates the signal emitted by that distrubance mentioned so long ago.

Just think, then – every time you download one of his albums off of bittorrent, 50 Cent has to do all that work. And you wondered why piracy was wrong.

I had this idea a while ago and never really developed it, so I place it here for further consideration. I think a lot about DRM software – that is, software designed to prevent piracy of information. It appears to me that attempting to build an impenetrable DRM system is akin to a special problem in encryption. Because there is plenty of literature and knowledge about encryption, converting the DRM ‘problem’ into a special case of encryption would greatly contribute to any consideration of a DRM method.

For the uninitiated, the basic encryption problem is that Alice wants to send a message to Bob, without anyone else, like Eve, being able to tell what that message is. The ‘DRM’ case isn’t too far from it – suppose Alice is a musician and Bob is a music buyer. Alice wants to be able to create a message such that bob can read it, but nobody else can read it, no matter how much information bob gives them. That is the goal of DRM software – so that only someone who has purchased (and therefore has license to use) some peice of copyright material is able to use it.

The thing is, though, that problem is completely absurd and no rational person would attempt it. Whenever Bob has read the message, all he’s got to do is give that message to Eve and the system is broken. You can gussy it up all you like with fancy equipment but as long as Bob is capable of comprehending the message Alice sends to him, he can just transmit that same message to Eve. Viewed as an encryption problem, DRM becomes entirely a fool’s pursuit. Something tells me the situation has to be more complicated than that, but I don’t really think it is.