Re: Executing Encrypted Code

[ph@netcom.com (Peter Hendrickson)]

At 5:25 PM 12/20/1996, Andrew Loewenstern wrote:
>Peter Hendrickson writes:
>>  I would expect software prices to drop because everybody using
>>  the software would be paying for it.

> I don't mean to sound rude or insult you personally, but this is utterly
> absurd.  If everyone is paying for the software then the company would be
> making even MORE money.  Only a fool would want to make less money!!  If
>money
> wasn't important to these people they wouldn't be in the business of selling
> software in the first place!

> Aside from putting a gun to people's heads, the only thing that lowers
>prices
> is competition.

Yes, I was assuming a free market.  Sorry if that was not clear.

Not only would software companies face competition in the decrypting-processor
market, but they also face competition with other packages running on
other platforms.

Presumably customers would need some sort of reason to use the
decrypting-processor.  Some companies may choose lower price.

> Only a fool would want to make less money!!

I agree completely!

>>  It is not out of the question for software vendors to sell
>>  two versions of the same software.  One is the piracy-free
>>  version and the other is the copy-as-much-as-you-can version.
>>  I would expect the piracy-free version to be substantially
>>  cheaper.

> That would render the entire scheme pointless.  It only takes _____ONE_____
> copy of the software to get out for the whole world to pirate it.

This was discussed a few posts back.

Let's say you manage to get the secret key out of the decrypting-processor.
That gives you the executable which could run on any decrypting-processor.
Since it is not authenticated (*) you can't run it on another
decrypting-processor.  You can run it in emulation someplace else,
but a heavy performance price is paid.  If the leading edge processors
are all decrypting-processors, a very heavy performance price is paid.
If the instruction set is kept secret, even writing an emulator could
become hard.

(* I am being inconsistent, incidentally.  At one point I said that software
would be authenticated once, but I now realize that to prevent multiple
uses it has to be authenticated for use on a particular processor, too.)

Peter Hendrickson
ph@netcom.com