Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Most “mandatory requirements” in corporations are imaginary

nibblestew.blogspot.com

Post reply on HN
Print view View original

401–405 of 405 posts

Re: Most “mandatory requirements” in corporations are imaginary

#401

Post by qznc »

BeetleB wrote: post #303
qznc wrote: post #138

Earlier quoted context omitted.

An example, I'm unable to resolve so far: ISO 26262 includes the requirement to prevent “implausible values, execution errors, division by zero, and errors in data flow and control flow” (8.4.4). I'm confident the division by zero aims at integer division because the result is undefined. For floating point, division-by-zero is perfectly fine and well defined by IEEE 754. Nevertheless, our safety folks require us to e…

I worked briefly in the ISO 26262 space. While I don't have a copy of it in front of me, my understanding was not so much that you must prevent division by zero, as much as it is that you handle division by zero appropriately . Appropriately means that you can show that if you get a divide by zero, safety is not compromised. If I get an Inf or a NaN in a calculation involving cruise control, and I do nothing to handl…

What I meant by "the same effect" is that division by zero or by a small number can both result in Inf.

You are correct that it is possible in general. However, in this specific project, the decision was to shut down the device for safety reasons if the CPU detects a floating point division by zero. In my opinion this reduces the availability of the functions unnecessarily.

Re: Most “mandatory requirements” in corporations are imaginary

#403

Post by colejohnson66 »

corty wrote: post #240
colejohnson66 wrote: post #214

Earlier quoted context omitted.

Well, it’s incorrect when dividing by 0, yes. But when you do the limit as the bottom factor approaches 0, it makes sense. So, below, the first one is defined, while the second one isn’t. lim_{x->0+} 1/x == +Infinity 1/0 != (anything)

Yes, but thats only a small part of the story, your limit only works if the numerator is constant, positive and > 0. Actually, the correct answer for lim{x->0} a/x is really "anything" and "strongly depends on a and da/dx"... So in general, assuming anything like \infty for a division by zero is wrong wrong wrong.

Sorry if I wasn’t clear. I was not implying that any a/b where b approaches 0 is always +infinity. I was simply saying that a/0 isn’t defined, but that in the case of 1/x (positive 0), it is somewhat defined as positive infinity.

Re: Most “mandatory requirements” in corporations are imaginary

#404

Post by bonzini »

qznc wrote: post #138
sfifs wrote: post #25

In BigCorps, if there's a stupid requirement, there's usually a reason for the stupid requirement to be there in the first place but getting to the reason might require un-peeling a few org layers to since the people enforcing the policy will not be the people who wrote the policy. A more productive use of time would be to understand the reason for the policy, document out why it doesn't apply to your case and then a…

An example, I'm unable to resolve so far: ISO 26262 includes the requirement to prevent “implausible values, execution errors, division by zero, and errors in data flow and control flow” (8.4.4). I'm confident the division by zero aims at integer division because the result is undefined. For floating point, division-by-zero is perfectly fine and well defined by IEEE 754. Nevertheless, our safety folks require us to e…

> This is stupid because a division by a very small number has the same effect as division by zero.

It doesn't, it cannot ever return NaN.

Re: Most “mandatory requirements” in corporations are imaginary

#405

Post by yebyen »

spiritplumber wrote: post #397
arethuza wrote: post #27

Earlier quoted context omitted.

Forget INTP, Sociopath|Clueless|Loser is definitely the most effective categorisation of roles in organisation I have encountered: https://www.ribbonfarm.com/the-gervais-principle/

This is really interesting, thanks for sharing it! Is this how American offices actually work?

I just learned about this apparently popular interpretation and after going out and reading about it enough, I had to come back and find this thread just to say:

"I'm in this photo and I don't like it".jpg

I'm afraid it might be true! This explains a lot for me.

Post reply on HN