←back to thread

171 points _sbl_ | 1 comments | | HN request time: 0.288s | source
Show context
oc1 ◴[] No.44522688[source]
<< According to official records, the design for the bridge shifted multiple times over the past seven years, largely due to conflicts between the Public Works Department (PWD) and the Railways. The two agencies couldn’t agree on how to share land, and in trying to work around both railway property and the new Metro line, they ended up producing a final layout with an abrupt 90-degree angle.

I love that mindset. Europeans would have simply refused and 100 years later it would have probably been build after all legal has been cleared. Indians instead never say no. That's how you build software, so why not bridges.

replies(14): >>44522771 #>>44522797 #>>44522901 #>>44523126 #>>44523197 #>>44523198 #>>44523260 #>>44523451 #>>44523656 #>>44523670 #>>44523959 #>>44527192 #>>44527686 #>>44529344 #
cs702 ◴[] No.44523260[source]
Shouldn't the bureaucracies be penalized, instead of the poor engineers?

The engineers built the 90-degree layout specified by their clients!

I wouldn't be surprised if there's a paper trail documenting the engineers' objections, signed and notarized by the clients.

It's hard for me to judge the engineers without knowing more.

replies(6): >>44523356 #>>44523517 #>>44523544 #>>44523630 #>>44525019 #>>44529354 #
Enginerrrd ◴[] No.44523544[source]
No. I'm a licensed civil engineer in the US. The license comes with an explicit duty to the public, to uphold public safety. I am in responsible charge of the work I produce and personally liable for the safety of that work, in perpetuity, and it SHOULD be that way. Any plans that I produce are subject to that standard.

India has a similar system for public works projects where a licensed engineer MUST supervise the work.

Frankly, sometimes I think the software world would be a lot better off with a similar system.

replies(2): >>44523746 #>>44529393 #
FinnKuhn ◴[] No.44523746[source]
This doesn't look unsafe though, just inconvenient.

Both in the US[1] and UK[2] you can find bridges with actual 90-degree angles. The one in India[3] is more like 75 degrees.

[1]https://maps.app.goo.gl/3CBqVHbVEtonHjcr9 [2]https://maps.app.goo.gl/8cVB44VDJRPadY6s6 [3]https://maps.app.goo.gl/ikPSmLEGYwVJLqDz7

replies(5): >>44523897 #>>44523958 #>>44524021 #>>44524696 #>>44525660 #
Enginerrrd ◴[] No.44525660[source]
Those are not the same at all.

Search for "swept path analysis" for just one component of what you're missing. (There are many other components of design of a curve like this to consider.)

A 90 degree change in direction is fine by itself provided there is sufficient radius for vehicles to make the turn at the design speed.

In this case, if its two lane you may not be so convinced of its safety when it's your loved one on a scooter who got hit by a bus which tracked over into the oncoming lane just to navigate the curve. Or if its a single lane, when they died on the ambulance which was stuck in traffic on the bridge because two vehicles are unable to pass and everyone behind them would need to backup in unison to sort out the resulting cluster.

But safety is only part of the duty to the public here. The bridge needs to function for its intended specification and if it fails to do so for basic engineering reasons, you absolutely have no business holding a license and signing off on public plans and indeed you would be disciplined or stripped of your license for something like this.

replies(1): >>44527822 #
freetinker ◴[] No.44527822[source]
An elegant solution to this problem would be to enforce a very low speed limit entering the turn.

“Sharp turn ahead, reduce speed to 5 km/h”

replies(2): >>44528012 #>>44530786 #
1. Enginerrrd ◴[] No.44528012[source]
Even if the swept path issues were not an issue here, that method has been tried and tested, doesn't work, and is considered bad design practice in the field of traffic engineering.

There's a great deal of evidenced backed engineering practices and solutions to draw from to solve issues like this. Unlike software in noncritical applications, you can't just pull things out of a hat and hope it works. People die if you do that.

You can't just say "here's how fast you can drive this road." You have to design the features of a road to calm traffic. If you design a road like a freeway and then put up a 15mph speed limit sign, people aren't going to just drive 15mph. Thats just not how traffic and people work. If you want people to go slower you should design traffic calming features into the road.

Its generally bad to have inconsistent design speed for different features of a road. And by bad I mean people will get killed at disproportionate rates. It happens all the time anyway for various reasons but not very often.