Re:Bootstrapping trust in compilers
by xbtcdev on Tuesday December 08, 2015 @07:49AM
(#10696023)
The trouble with KTH is that somebody could go through all of this song and dance, and it still wouldn't mean a thing because how do you trust them?
[ Parent
| on HN ]
Re:Bootstrapping trust in compilers
by rejschaap on Wednesday December 09, 2015 @01:51AM
(#10702303)
You don't trust them. The result of this endeavor is not a trustworthy compiler, the result is a procedure to generate one. Every step in the procedure can and should be verified independently. What this buys you is a procedure that produces a trustworthy compiler given your initial environment is trustworthy. The latter still being an issue of course.