I wounldn't say it is unrecognized; it falls into the area of self-replicated automata. As such, it is related, however distantly, to
cellular automata (
The Game of Life,
linear tapestry automata, etc.), a traditional area of study with interesting implications regarding computability (GoL is in fact a Turing-complete computation system, as it can be used to
simulate a UTM), chaos theory, cryptography, and optimization theory, among other things.
If you want to experiment with self-replicating systems as a hobby, well, I'd do so under very specific conditions: to wit, I would get into playing
CoreWars and write my designs for that. It's clean, easier to work with that a real machine, and there's no risk of accidentally releasing something with unpredictable side effects (e.g., the
RTM worm, which has a minor bug in it that caused each dupicate copy to replicate out of control instead of suiciding) out onto the Net.
I would be very wary of programming live viruses and worms for real machines: the Richard Morris example shows that a worm or virus can be dangerous even without a payload. This is especially risky today, when the US government is in a paranoid frenzy over 'hackers', seeing threats that don't exist everywhere while often ignoring the real ones. Currently propsed legislation in many US States (i.e., the UCITA) would make it illegal to even own tools that could
potentially be used to retro-engineer software (such as, say, DEBUG.EXE or gdb - betcha didn't know you were in possession of an illegal cracking tool, did you?). Given certain clauses of the DMCA, ven if you never crack a program or write a virus which gets loose, you are putting youself into legal jeopardy just by studying exploits and system flaws, even if your intention is to report the bugs in question - at least one bug-hunter has been suppressed in this manner already (at least I recall such a case being reported, but I can't seem to find the link). Be careful with any experiments you go ahead with.