WHMurray@DOCKMASTER.NCSC.MIL (04/06/90)
>There is NO evidence that anything other than an .exe, .com, .ovl, or >.sys file can infect. There has been talk about .pgm files (for >dBase) and lotus spreadsheets being carriers but I have no evidence of >any known. "Not so, Mr. Schuyler!" That is a very large NO, and I do not wish to get into a shouting match with my learned colleague. Neither do I wish for the rest of you to be mislead. First, I think that my colleague speaks in the very narrow sense of MS-DOS. While this is the important territory for the moment, it is not all there is. Ken Thompson, in his Turing Award acceptance, describes a very credible scenario of a virus in source code. He makes the relevant point. "One man's data is another man's program." There is very credible evidence that a 1-2-3 .wks file, which looks for all the world like a data file, can contain a macro which will create a copy of itself in all .wks files such that use of those will cause further copies. That sounds like a virus to me. It is not clear that such a .wks file could achieve the feat of infecting .exe or .com files. And it is clear that it can only execute in a 1-2-3 environment. It cannot operate in a DOS, UNIX, or primitive hardware environment. But that was a BIG no, we do not wish to mislead, One man's data is another's program, and not everyone operates in a DOS only environment. My point is that anyone can tell a lie a in any language. Whenever you accept ANY data from another, you run some risk of being duped. While it is true that the virus writer must solve the problem of getting his program in control, and that that problem is more easily solved in some environments than in others, do not under-estimate the ingenuity of the malicious. While the current battle is being waged on the PC battlefield, and while the war may even be won or lost here, the phenomenon should be known in the broadest and most general light. William Hugh Murray 21 Locust Avenue, Suite 2D, New Canaan, Connecticut 06840