Saturday, July 18, 2009

Comparing Technologies

I heard that Wolfram Alpha was finally available, and I wanted to try it out. Wolfram Alpha is designed to be more than a search engine -- it's an answer engine. A search engine tries to find Web documents that contain information you want. But Wolfram Alpha will try to calculate an answer for you from data that it can access.

For example, if you want to know the "weight of the earth in pounds", it figures that (1) by "weight" you really meant mass, (2) the earth mass is available in a table of data about the planets of the solar system (although in metric units), (3) a table of conversion factors is available, and (4) a formula for converting units is available. Moreover, it has the 'smarts' to know that this is the data needed to get the answer, and it knows how to find and combine the details to get the answer.

Now, what problem would I use to try out this new answer engine? Well, I recall reading that DNA is an incredibly dense data storage and retrieval system, but I didn't have any number for the data density in, say, bytes per pound. So, I tried to get the number from Wolfram Alpha. But "DNA in pounds" was not precise enough. How much DNA? Just one 'base pair' (one unit of the chain), or an entire chromosome? And if a chromosome, which kind? (because they have different lengths)

DNA is a chain of information units called nucleotides. The chain is shaped like a twisted ladder, with each rung a pair of nucleotides that encodes two bits of information. There are four kinds of the nucleotides, so I began by asking for the mass of each kind, using their chemical names:

adenine mass in pounds: 4.9468*10-25 lb
guanine mass in pounds: 5.53252*10-25 lb
thymine mass in pounds: 4.51683*10-25 lb
cytosine mass in pounds: 4.06729*10-25 lb

I also needed the mass of the 'backbone' unit, for the 'sides' of the ladder:

deoxyribose mass in pounds: 4.45458*10-25 lb

Then, assuming that the four nucleotide types are used equally, I could now compute the data density of DNA:

1.084547*1024 bytes per pound
(That's about a one followed by 24 zeros.)

Now, what man-made data storage and retrieval system could I compare this to? I have an 8 GB thumb drive that weighs a quarter of an ounce, which may not be the most dense, but it's denser than a DVD or a hard drive. I calculated it's data density to be:

5.5*1011 bytes per pound

That means that DNA is about two trillion times more dense than the thumb drive. That is, the data capacity of a quarter of an ounce of DNA is equal to about two trillion 8 GB thumb drives! Engineers would love to be able to design a data storage and retrieval system with the density of DNA, but they don't know how.

Yet there are atheistic scientists that believe that mindless evolution accidentally created DNA millions of years ago. I have two reactions to this evolutionary belief:

First, as an engineer, I feel insulted that people actually think that a random process can out-do what none of my engineering colleagues can accomplish.

Second, it is clear to me that I don't have enough faith to be an atheist.