01 July 2005

Information is information

Topic:

Scientists at Chiron have developed a new type of vaccine. It protects babies from catching a common herpes virus from their mothers during birth.

A major difficulty in creating the vaccine was that the virus exists in many different strains. A traditional vaccine, made from the virus itself, would not protect against all of the strains.

So the researchers looked at the genes the strains had in common, and made a vaccine from a subset of those common genes. In other words, they looked for identifying characteristics of the virus.

Interestingly, this is exactly how anti-virus programs on computers work — they use a virus signature, which tries to describe the essence of a virus, the parts which cannot be changed without deactivating the virus. When that signature is found in a computer system, the virus can be eliminated.

The Trustworthy Computing (TC) initiative, spearheaded by Microsoft, is an attempt to take the opposite approach. They are trying to capture the essence of a computer system — its DNA, if you willl — and will then regard any deviation from that signature as an intruder to be eliminated.

TC will make systems safer from attack. Unfortunately, it will also make them hostile to innovation. Let's look at another example from biology.

The 18 June issue of The Economist contains a story about parts of our DNA called retrotransposons, or "jumping genes". These look very similar to retroviruses. I think it's not too far a stretch to suppose that at one time they actually were free-living viruses, which conferred enough benefit to their hosts that they eventually got incorporated into our genome.

And what benefit was that? Greater brain complexity. These "jumping genes" make it more likely that the precursors to brain cells would turn into neurons rather than the other kinds of brain cells. They also make the neurons more diverse, providing the raw material for our intelligence.

So it's possible that we became human because our ancestors were infected with a virus. If their bodies had had the biological equivalent of TC, homo sapiens might never have developed. I fear that with TC, computing will not be able to develop.

Biology has been learning many lessons from computer science. It's time for computer scientists to listen to biologists, too.