> That's like saying you deliberately pulled the fire alarm when there's no fire. Yes, the system worked as designed. The people got out and the fire trucks arrived.

So - this is a good point! Honestly this is what I think we should be debating: "was it right or wrong to pull the proverbial fire alarm?"

That's why I've stuck so hard to my point - if we inaccurately compare it to malware, we never even get to these interesting ethical questions (we'd instead be stuck in a shouting match about "it's bad because malware is bad!" ).

> I'm willing to bet that from now on it will require a two-person system to change anything like this again, because of the abuse. And that makes the system worse.

You're probably right about the effects - management has a tendency to lock down things when this happens. I think that's tragic, but I don't personally lay all the blame at her feet for it. Google management is ultimately capable of making a different decision, and they do not need to take that course of action.

My gut tells me that they will take that course, based on what I've seen management do at other companies. I guess I just have to point out that their response is a choice, and it could be different.

> It's similar to imagine if schools required two people to independently pull the fire alarm at the same time in order for it to go off.

In a way, yes. But it also shows something else - schools would never actually do that, because it's absurd and makes things more dangerous. Instead, a school would focus on "why did this person pull the fire alarm in the first place" and address that instead of locking down the fire alarm.

Google management could do the same, in our analogy. But... that's a lot harder, so I imagine they'll take the easier route.

> Absolutely not, no. Again my analogy about how someone could replace the frontpage of google to no longer have a search box, only a copy of the US constitution. How could this POSSIBLY be about the truthfulness of the message?

I apologize - I think I wasn't clear enough with my words. I meant "right or wrong" in the sense of "ethically right or ethically wrong," not in the sense of "is it factually correct or factually incorrect."

> If you don't think this kind of change involves risk

I didn't say it didn't involve risk. I was saying that it's a "slippery slope" argument because we have no evidence to suggest that she didn't take the risks into account, and we don't really have any evidence that the change she made could have realistically led to the outcomes you mentioned. We have to fill in a whole lot of blanks before those outcomes seem probable: and hence why I called it out as a slippery slope. Not because it's completely impossible, but just that it doesn't seem probable based on what we know.

> then I'm sorry, you don't understand complex systems.

I know sometimes comments get heated; and I know I've sometimes said things like this in the past. But that's just really unkind. You and I don't know each other, and I don't think it's appropriate to insult each other's intelligence like that. If you're interested in my career history and want to know about the complex systems I've worked on, I'm happy to share it.

But insulting me in your argument isn't appropriate. If you don't want to discuss this stuff further, that's fine. No need to be needlessly cruel.

I didn't mean to insult. I saw your comment like the defense "sure, she was driving drunk. But nobody actually got hurt!". I think it's fair (and generally for drunk driving it is the case) to treat the action at least partially from the potential harm, that was avoided due to mere luck.

I don't think it's a slippery slope to point out that it could have had unintended side effects in addition to the (for Google unwanted) intended effects.

And unlike with drunk driving, complex systems break every day from honest "oh wow, I was sure this couldn't cause an outage".

When someone has honest intentions there's no blame[1], but when someone doesn't, then yeah they are to blame for any outage it caused, and for damage that statistically would happen for every N times someone did this.

[1] including if someone bypasses an 'annoying' safety feature. Because it should be set up such that the safest way is the easiest way.