Hacker News Reader

Comment by chromoblob | Hacker News Reader

The program must not show bugs when run on a hardware with unforeseeable bugs, you call this reasonable?

And yet that's what every good engineer did when Spectre came out. Same with the Pentium fdiv bugs, and same with a host of microcode bugs that come up all the time.

Not my business to decide what you think is reasonable. That's just what happens in the world, and what (in my view) good engineers sign up for.

The choice is between letting hardware be not finalized and letting that force software to be non-finalizable, and letting software be finalizable and forcing the hardware to be finalized too. I like latter more. Finalized hardware is better by itself as well.

We would all like bug-free hardware, but we won't get it and our job is to write good software in the environment we were given

> we won't get it

Why do you think so?

> The choice

What choice? I have to fix bugs today as they come.

If you buy a car and the airbags randomly deploy, would you consider it reasonable for the manufacturer to respond "oh, yeah, that'll happen if you drive it on roads rougher than polished stainless steel. You should only be driving on polished roadways"?

If this requirement was known to me before I bought, sure.

I think that this is a bad analogy to hardware, though. Polished steel roads are unreasonable to ask for, but bugless processors are reasonable to ask for.

No, they aren't. You can only buy the buggy processors that exist, not notional bugless ones.

As if processors were only sold by alien gods, and processors without bugs were impossible to create.

They may not be impossible to make, but they don't get made. And I guess you think Microsoft has been putting bugs in Windows for the last few decades just for something to do, or was that also the influence of alien gods?

Reply on news.ycombinator.com