|[4 earlier articles]|
|Re: code transformations? email@example.com (Darius Blasband) (1996-09-26)|
|Re: code transformations? firstname.lastname@example.org (1996-09-26)|
|Re: code transformations? email@example.com (1996-09-26)|
|Re: code transformations? firstname.lastname@example.org (1996-09-29)|
|Re: code transformations? email@example.com (Norman Culver) (1996-09-29)|
|Re: code transformations? firstname.lastname@example.org (1996-10-03)|
|Re: code transformations? email@example.com (Henry Dan Lambright) (1996-10-20)|
|Re: code transformations? hogan@rintintin.Colorado.EDU (1996-10-24)|
|Re: code transformations? firstname.lastname@example.org (1996-10-24)|
|Re: code transformations email@example.com (Henry Dan Lambright) (1996-10-25)|
|From:||Henry Dan Lambright <firstname.lastname@example.org>|
|Date:||20 Oct 1996 16:52:51 -0400|
On 29 Sep 1996, Torben AEgidius Mogensen wrote:
> Tom Lord wrote:
> ... This got me to thinking -- suitably
> sophisticated transformations could preserve the semantics of the
> program, while completely disguising the text of the source. This has
> obvious implications for anyone into intellectual property theft.
Suppose you modify a program's source, but do not modify it in a way so
that it performs differently (ie the modified version never yields
different output for any given input.) I would like to know if the
underlying control and data flow graphs would be similar enough (between
the former and later versions of the program) for you to be able to "diff"
the two graphs. ie. the text may have been changed, but the underlying
functionality (represented by the graphs) remains the same. Does anyone
out there know if an algorithm exists which counts the minimum number of
steps to change one DAG into another ?
Return to the
Search the comp.compilers archives again.