| Fermilab |
Home |
Pillars |
Channeling | Adv. Accel. | Infrared/Dyson | SETI Home |
Biography | Bibliography | SETI Hacker | Student |
| Introduction I welcome comments on the SETI Hacker conjecture particularly by credible experts who identify themselves and their affiliation and can explain their views quantitatively and with references to serious literature in a way I can follow. Comments are easier to digest if they don’t start with some reference to a work of fiction, a comment about Microsoft or Apple, or “you idiot.” Typically I do not respond to individuals but update the material below to reflect some of the critical comment. I have not yet found one computer expert who believes it is possible to hack a computer without knowing something about its inner workings. They all think I am silly (at best). I have been surprised that in the face of such a consensus no technically-gifted contrarian has come forward and said “here’s a way.” I am not a computer expert. I got my computing start on Illiac I (pre HAL in the SF metaphor) using the manual noted at the Sydney Illiac site. From the late fifties through the eighties I was involved with on-line work and large computers. I have since become dated. I have handled the analysis for my last two research projects using large Excel spread sheets. This can work for an astronomy investigation involving several hundred thousand sources, not for particles physics experiments. |
Two branches to the SETI Hacker conjectureIn the Acta Astronautica article there are two different branches for the SETI Hacker conjecture. To quote the article "At least two scenarios need to be considered in protecting against a malevolent SETI Hacker signal. One is a computer virus in the message that takes over the computer at the receiver. The other is an open message that gives an impenetrable software code or instructions for a hardware translator to handle an opaque message." Most comments have addressed the explicit SETI Hacker branch, where a virus downloads directly into the computer without human involvement. The second scenario may be the more plausible possibility particularly in view of the serious criticisms of the first branch. Few of the comments seem to consider this branch and fewer still have questioned it. topMost significant criticismsTo enter a computer the virus must run but to run the virus must enter the computer. Put differently “If the software does not run it can't untangle inner workings. If it can't untangle inner workings it cannot run.” This is the bootstrap problem that was mentioned in the article. It is a serious, perhaps the most serious, criticism of the explicit SETI Hacker conjecture.All computer viruses work by
exploiting
known systems and
architectures As noted above, this view seems to be
universally held by computer scientists. I have been surprised by this
comment. In biological systems this does not seem to be necessary. On
the other hand biological viruses have selectively evolved for
penetration. Must know exact form of
instructions and all other material This is true for most earth computer
virus situations. The SETI Hacker conjecture suggests the possibility
of a small kernel that would somehow booststrap in, use the kernel to
investigate the internal computer code and hardware and then fomulate
code. This is a very big jump beyond what most experts think is
possible. Only get one chance to penetrate
the host If the kernel is stored in the computer and bootstraps in it could have repeated chances. top ModelingModeling the SETI Hacker – earth-based computer relationship Marcus Leech has prepared a thoughtful, many would say damning, critique of the explicit SETI Hacker conjecture. I have attached my annotated version. I understand that Leech is continuing to work on his model. I regard Leech's work as important. Further work on this could help to illuminate some of the factors that enter into this subject. Several obvious factors are the size of the instruction set and memory size. Leech has also drawn my attention to the International Obfuscated C Coding Competition. This competition is interesting both from the standpoint of SETI message size and the SETI Hacker conjecture. The contest is doing research on SETI messaging!My Fermilab colleague, Bill Higgins,
suggests “We ourselves
could provide a 1940s-level civilization with instructions on how to
build
modern computers, and how to bootstrap manufacturing, materials
science, etc.
to do so. But we couldn't guarantee that their computers wouldn't
crash.” top SETI Hacker MotivationWhat
about motivation? In
discussing SETI signals SETI astronomers talk about altruistic signals.
The
problem is that there is a strong motivation to spread a culture
through
one-way
electromagnetic signals. It is the fastest way to send intellectual
spore
across the cosmos. Most spores care little about the soil they settle
in. They
just want to propagate and grow. On balance, my view is that most
signaling
will be self-centered and one way. In that picture any signal that is
not just
cultural noise is probably antithetical to our situation since it wants
to get
embedded in the new world. Think parasites, the relation of the Spanish
with
the Incas, and on and on. top |
Entering a computer and runningModern computers are very complex and idiosyncratic. How could an internal virus find its way through the mess?Computers may be messy but there are some general principles that were recognized early in their development. Most additions beyond that have a basis in logic. My understanding is that in many contemporary computers allocation of data and program areas is handled by the operating system. Does this mean that during failures such as soft power failures some of these rules may be violated? Wikipedia contains a short discussion of this. For additional information see an article on stack overflow noted by Leech. top A SETI Hacker program will be difficult to untangleInstruction set It appears to me that over a fifty year period instruction sets have been somewhat invariant. There’s a good reason for that, they are dealing in a universal language, arithmetic/logic with a relatively limited number of operations. Byte size (using byte generically) Some comments along this line are nicely abstracted in a quote from Daniel Joyce “Aliens can't even be sure we use trinary, or binary encoding, or some other form. Do we use network byte order, host byte order, are our numbers 1 or 2 s complement? What's the ISA? Intel, MIPS, Sparc, or Rigelean SP500? What if it's not even a 'normal' processor, but one tied to running a particular language well, like a LISP machine, or the old-school weird APL ones? How is their OS even supposed to start?” For Joyce's list the combinatorial total might be O(100) - O(10^5). This is no challenge to untangle to a code breaking or Google-like system. topSETI Hacker conjecture as fictionNancy Carrigan and I published some SETI Hacker fiction many years ago in Analog, “The Siren Stars”. Used versions can be picked up at Amazon.com. A beautiful illustration of the Analog magazine cover by Kelly Freas featuring the serial version in March, 1970 is available. We used fiction because I did not see any way to handle discussion of the SETI Hacker conjecture in the main stream science literature. Other works mentioned in connection with the SETI Hacker conjecture include Hoyle and Elliot's 1962 novel A for Andromeda, Sagan's 1985 Contact, Star Trek: The Next Generation Story, Contagion,Virus, and Independence Day.Parenthetically, my daughter, Caroline Carrigan, was Scotty's ex-wife (Pamela Denberg Doohan) in the celebrated December 20, 1986 Saturday Night Live spoof of Star Trek that contained Shatner’s famous refrain to the assembled Trekkies “get a life”. Many SETI Hacker critics have made the same suggestion to me. top Interesting comments and observationsIt is actually easy to hang up calculators, for example, a division by zero in an iterative sequence is typically not trapped and will hang up the sequence. Maybe some of the SETI Hacker critics should also get a life. top |