On Thu, Jul 30, 2015 at 1:04 PM, The Doctor <drwho@virtadpt.net> wrote:
On 07/30/2015 01:32 AM, oshwm wrote:
would a pre-internet era set of IC's and components kick-start the process a little without losing too much trust? Use slow old IC's in parallel to gain something usable :)
Something like this? http://cpuville.com/ http://www.homebrewcpu.com/ Or maybe something like this, seeing as how we really can't trust anything integrated as the microscopic level or smaller? http://6502.org/users/dieter/mt15/mt15.htm
Those are close and would certainly be a goalpost in the rebuild. You might be able to trust logic gates because you can exhaustively test their logic. On the other hand, how do you know that once you connect enough of them to each other that their secret gates inside don't sense each other and activate? Since you're that close to stone age anyway, why not start one more step back at relays and core memory. It's like trying to validate a 256 bit blackbox hash function someone gives you... sure, their supplied test vectors may all pass, but you have no idea or way to test what happens when you start pushing real data through the secret instructions. You simply can't test all the possible data combinations so you have to throw their box back in the snakeoil.
Hell, why don't we just start building DCPU-16's and bootstrap from there? https://en.wikipedia.org/wiki/0x10c
That's too complex, and the first one pre-exists so it's chicken and egg.