Hacker News Reader

Comment by kzhahou | Hacker News Reader

> Maybe he just deployed a webapp that the people in the communications center could access...

I can hardly thing of a single bit of code that should be deployed in such a literal life-and-death environment, which shouldn't go through FULL REVIEW.

Maybe it's a small webapp that gives a bit more information -- which leads dispatchers in the wrong direction one out of ten times. Maybe his "highlighting the most critical calls" was highlighting the wrong ones. Maybe he solved a minor problem only, but took up dispatchers' time in training.

View on HN

Clemens is in the comment thread. It seems like the system he wrote read out of the 911 call database and displayed all the active calls. Then highlighted calls based on keywords so the head dispatcher could more easily see high priority calls ( crimes in progress ? ).

This seems very safe. The captain in question is just protecting his turf. Even if he saw the benefit he would never admit it because it wasn't his idea.

They hired Clemans as a programmer and already implemented some of his ideas - I don't think jealousy is a motivating factor here.

I'd love to read an 'AMA' style post, or even just an interview done by another programmer, to learn more about what exactly he did, and how he did it.

Phrased more charitably, "protecting his turf" is actually "doing his job."

With the little information available, I can see the captain's position. The programmer was new and rolled out changes to the surprise of the captain. Programmer hadn't earned captain's trust yet, and was pushing out new procedures which the captain did not (?) agree with. That'd be a no-no in any company, just substitute CEO for captain.

Regardless of whether his system was fantastic, the manager in any org is meant to be a safeguard, even when it feels like a bottleneck.

I wasn't in trouble for doing what I did. The meeting was discussing how to keep the innovation going in a more controlled manner. The Captain believes the only way to implement software in 911 is to take years and millions of dollars to do it.

How is that very safe? Presumably were looking at this unreviewed web app instead of what they were looking at before.

The dispatchers I observe use it focused on their primary screens and if something highlighted they saw out of the corner of their eye.

How about all the important calls they let languish because of the bide-shedders interfering? How is that safe?

Can the naysayers be sued by the family of some kid who died because dispatchers weren't allowed to evaluate a new tool because change scares some people? (No, of course not. And that's why there are so many bikeshedders.)

Bikeshedding is when leaders ignore complex & core questions to focus on trivial ones that are easier to understand. The tools that dispatchers use to handle incoming called doesn't sound like a trivial question.

The core of bikeshedding is continued nitpicking at the trivial by those who can't understand or don't know enough to grasp the its actual significance. And it's not just leaders.

Certainly change would be bad if the programmer in question blidnly modified existing systems. But I don't see any of the reactionaries asking that - they just assume the worst and go off on how unsafe it might have been.

But a new webapp bundling data from other systems into a new view? It shouldn't be any problem at all. The professionals on the phones are the very ones who would have to evaluate the information it returned so there's no way around letting them use it.

The code itself might be intrinsically safe, but the process it generates might not be. The argument doesn't have to be "this code will crash the 911 system"; it might just be "people are going to get hurt when operators act on the information it attempts to generate".

And yet if some operator had an idea to keep some notes (license plates from APBs or something) on stickies and pass them around the room - a new procedure - nobody would act like the world would end.

"FULL REVIEW". How FULL exactly? I think a really thorough review should involve the people who defined the language used, and who wrote the compiler. Certainly everyone who ever worked on the code. And the Compliance team. Lets involve them early to speed up the process!

The professionals on the phone seem to have spoken very highly of the benefits from this program. Why are people second-guessing them?

Maybe it did go thru full review. I can't tell from the post whether it did and the police chief just decided post-hoc that no changes could be made without their express approval (i.e. in the future), or whether Clemans bypassed existing protocols.

And sure, maybe his changes were imperfect. But surely lots of changes made by others there could have had terrible consequences.

Or maybe his "highlighting the most critical calls" was simply visualizing info that the dispatch personnel were already entering.

Still, I can't tell whether he did anything wrong from the post itself. And I guess I'm inclined to believe he didn't, but maybe my prior belief is inaccurate.

If the chief/captain/whatever didn't know about it, it wasn't reviewed. Simple as that.

Reply on news.ycombinator.com