Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!mcsun!sunic!draken!bjornl
From: bjornl@tds.kth.se (Bj|rn Lisper)
Newsgroups: comp.arch
Subject: Re: Self-modifying code
Message-ID: <BJORNL.89Oct11160723@octopus.tds.kth.se>
Date: 11 Oct 89 15:07:24 GMT
References: <1080@mipos3.intel.com> <48682@ricerca.UUCP> <236@bbxsda.UUCP>
Sender: news@nada.kth.se
Organization: The Royal Inst. of Technology (KTH), Stockholm, Sweden.
Lines: 11
In-reply-to: scott@bbxsda.UUCP's message of 10 Oct 89 22:03:57 GMT

Self-modifying code can be seen as partial evaluation delayed to runtime.
Partial evaluation, in its cleanest form, is to use the knowledge of some
function arguments to simplify the function. Consider a function f(x,y) in
two arguments x,y. If we know the value of x, say that x=c, then we can
substitute c for x in the call f(c,y) and obtain a simplified "residual"
function fc(y). Usually this is done at compile-time. A prime example is
substitution of symbolic constants. Nothing prevents, however, that this is
done at runtime (if, for instance, x is not known before then), which then
results in self-modifying code.

Bjorn Lisper