Zukoski1@hypermail.apple.com (Peter Zukoski) (11/22/89)
Dear virus-folk: thanks for all the responses to Richard Dawkins questions. Here's some further thoughts from Richard on the topic of mind viruses...He and I would be interested in your opinions, especially on evolving/mutating virus technology. Has anyone seen viri which evolve, or mutate in response to the environment which it is in? Or viri which recognize and "use" other viri which might be present? OK Richard, you may begin... - ---------------------------- There is something important about the distinction between what I call Anarchic Replicators and Socialized Replicators. In the world of DNA, anarchic replicators are things like viruses, or smaller units with names like viroids or plasmids, which parasitically exploit the large-scale transcribing and copying machinery in cells, machinery which has been put together by cooperating teams of socialized replicators. Socialized replicators are the ordinary mainline genes that travel from generation to generation in sperms or eggs and cooperate to build big survival-machines like our bodies. In a sense, our 'own' genes are parasitic on each aother's efforts, just as virus genes are parasitic on the efforts of other genes. In the world of mind viruses, the reason large religions like Islam or Roman Catholicism fascinate (as well as repel) me is that they are large aggregations of mutually socialized viruses, which work to sustain other members of the cluster and work to destroy alternative replicators. e.g. Islam has a rule that apostates must be punished by death. The rule of priestly celibacy in Catholicism at first sight doesn't seem like a self-preserving replicator. But it frees the priest's time for more active proselytizing, and it enjoys a good mutualistic relationship with another rule of contraception among non-priests. In the world of computer viruses, I find it harder to find an analogy for socialized replicators. I gather that anti-virus programmers have already used the 'biological control' self-replication technique - sending in a tame,'good' virus to catch the bad one. This reminds us of the possibility of a kind of ecology of computer viruses building. We are reminded of this, again, in another of the papers which states that viruses that were originally intended to be benign can turn unintentionally malignant when the user upgrades to a new Operating System. A given OS serves as an environment in which a virus may flourish or not, may behave benignly or malignantly. Couldn't we envisage a time when the whole computer environment facing a new virus is put together, not just by one monolithic OS, but by the OS plus a motley collection of aleady-infiltrated viruses, some benign, some malignant, s ome medicinal. New viruses will be written to flourish, not just in known OSs, like System 6, System 7 and so on, but in a background containing an unknown but statistically guessable collection of already existing viruses. Already, the environment that even a legitimate programmer has to cope with is more than just the operating system - think of the motley collection of co-resident INITs and Desk Accessories that you have to worry about when writing a program for public consumption. A virus ecology will just complicate the picture, in the same 'biological' direction. The language I am now using is not too far from the language that I use (e.g. in The Selfish Gene and The Extended Phenotype) to describe the co-evolution of genes in genomes. So far, as far as I know computer viruses don't evolve. Even with viruses manufactured by conscious human programmers, something like a mutually co-evolved cluster of viruses could come to constitute the environment to which any future virus has to accommodate itself. If truly evolving (self-modifying in adaptive directions) viruses start to arise, the trend will go even further. Software Companies may have to write their products to be compatible, not just with numbered versions of 'The System', but with the changing statistical ensemble of fellow-travellers both good and bad. I'd be interested in hearing whether any of this makes sense. Richard - ----------------------- "Do what you want -- you will anyway." PeterZ