From: Nick Tarleton (nickptar@gmail.com)
Date: Wed Jul 16 2008 - 13:38:48 MDT
On Wed, Jul 16, 2008 at 11:13 AM, Stuart Armstrong <
dragondreaming@googlemail.com> wrote:
> >> Actually, the difficulty I had in mind was the seeming impossibility of
> >> *proving* one's source code to another. Sure, one SI can just send her
> >> source code to another, but how does the other SI know that it's not
> just
> >> some source code that she wants him to think she's running?
> >
> > A trusted third party with appropriate access (not necessarily a mind,
> maybe
> > just the operating system) could confirm this.
>
> Still don't see how that could overcome an SI deliberately intended to
> deceive. Commands such as "do this for ten thousand years, then change
> your code to this" would not be detectable.
If such commands were obfuscated, of course they could be a problem, as
could any number of other obfuscated commands.
> You'd need to give it a
> very thourough check, be in control of the hardware while doing so,
> and have an idea of the history of the source code.
>
The trusted third party could be in control of and inform you about the
hardware.
http://en.wikipedia.org/wiki/Trusted_Computing#Remote_attestation
I don't see why the history would be necessary, although knowing that a
hostile agent tampered with the source could make you more careful of
obfuscated nastiness. Either way, though, you'd want to rigorously prove
cooperation if possible.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT