VX Heaven

Library Collection Sources Engines Constructors Simulators Utilities Links Forum

Computer viruses (BMJ 299-66a)

John Croall, Ian McKay
British Medical Journal, vol. 297, pp.981-982
ISSN 0959-8138
October 1988

PDFDownload PDF (343.88Kb) (You need to be registered on forum)
[Back to index] [Comments]
\text{T_EX size}

Dr Patrick J R Harkin (10 September, p 688) doubts the hypothesis described by one of us (JC) (13 August, p 488) that computer viruses may evolve (and may already have evolved) by random substitution and cumulative selection from preexisting pieces of software.

Firstly, the hypothesis is not that a computer virus could spring into being from nowhere by random chance alone, as Dr Harkin assumes, but that it could evolve from a pre-existing piece of software in its host computer in the same way that human viruses probably evolved from a host organism's DNA. Our study of viruses in the computers that use the CP/M operating system has led us to believe that a specific two byte point mutation occurring in memory at the right place at the right time might be enough to start a spontaneous virus on its way.

Secondly, within some long established computer networks, in which evolution has had plenty of time to occur, the phenomenon of "ghosts" is well known. Ghosts are fragments of old, deleted programs that have accidentally attached themselves to other programs and occasionally become activated by accident. Ghosts and viruses differ only in their kinetics of transmission.

Thirdly, Dr Harkin estimates that, in the current industry standard microcomputers - for example, IBM personal computers - that use the MS-DOS operating system, the minimum size for a computer virus is 97 bytes. Following his example, one of us created a virus this morning. From the initial idea it took only 20 minutes of work and the virus was up and running and infectious, capable of spreading spontaneously from disk to disk. Its minimal size for infectious behaviour was only 32 bytes; to ensure minimal pathogenesis and therefore probably long term survival a more sophisticated version of about 64 bytes was necessary.

Finally, may we point out that Dr Harkin himself is the result of an extremely unlikely event. The odds against his creation by the normal random reshuffling of his parents' chromosomes are one in 2^46 that is, one in 70 368 744 177 664. Yet that, we presume, is exactly what happened.

John Croall
Public Health Laboratory,
William Harvey Hospital, Wilksborough,
Ashford, Kent TN24 OLZ

Ian C McKay
Department of Bacteriology and Immunology,
University of Glasgow, Glasgow

By accessing, viewing, downloading or otherwise using this content you agree to be bound by the Terms of Use! aka