Development quote of the week

Bridges are built on time and budget by being "expensive" in time and materials. Yes they could be made more expensive, but they could also be made a lot cheaper. In fact in the hey day of road/bridge building of the late 18th and most of the 19th century they were built that way. And bridges failed a LOT. Enough that things like the https://en.wikipedia.org/wiki/Tay_Bridge_disaster were big, they weren't remarkable as small bridges failed weekly. What changed was that bridge failures started costing groups of people more money than they did when 1 or 2 people died on them. The groups which insure businesses find that to be expensive to their interests and they start putting in more costs for their policies unless things are changed. That causes industries to start enforcing (usually via government policies) engineering standards, a LOT of regulation and assuming that everything had to standardize on certain tools, materials, inspections, and other things. Those changes took a lot of social effort with many 'engineers' of the time calling them un-needed and pretty much whatever standard 'the government/industry/insurance has no business in making me do something different' excuses at the time.

[This isn't to say that the various regulations were good.. those took time to work out and to go from being overly cautious and too expensive to finding a 'reasonable' path. The big issue is that we all want a quick fix and the Universe doesn't care about our wants.. things take time and any regulation being from self-regulation to insurance regulation to government will take years to work out from being onerous to 'its just what we do']

So it isn't that we haven't invented the arch.. we just don't use it as much as we should because it take too much time, effort and just looks boooorrrrring.

> I understand the meaning of the example, once built the bridge is stable and will last well beyond its expected life

Except that that only applies to bridges that are similar to other bridges both in design and size, where things are well understood.

Someone mentioned the Tay Bridge disaster - built using a new material called cast iron.

Someone else mentioned Tacoma - the disaster wasn't (primarily) the wind and resonance. That was the proximate cause but there was a deeper failure which led to that. Despite resonance the Millenium Bridge was perfectly safe - just not a nice experience (which various engineer-trained people spotted).

And an interesting follow-on - I think it was the Tay Bridge disaster, which led to the Forth Rail Bridge being massively over-engineered, and its companion bridge the Forth Road Bridge is now being urgently replaced because the engineers have identified an unexpected failure mode - they think it could become unsafe even before the replacement is complete which will be a major disaster ... basically salt water has got into the sheathing that's supposed to protect the suspension cables, and the individual strands are now rusting and snapping - will the cables be weakened and snap?

Bit like the Titanic - if the water had been a bit warmer she would have survived the collision (yes I know, the iceberg would have melted and there wouldn't have been a collision :-) but she was built of that new(ish) material called steel. And what the engineers didn't realise was that in cold Arctic waters the steel became brittle so the iceberg ripped a big gash. If it had been a bit warmer it would have made a massive dent, but not breached the hull.

At the end of the day, it takes a while for cutting edge engineering to become safe and boring :-)

Cheers,
Wol

Right, people are usually paid to design bridges, and not via "please sponsor me" links on github.

Agreed. Regulation is what makes bridge building both a predictable process and product. Bridge designers have to be certified as do the builders of key components. Much bridge maintenance can be performed by your average person in a county maintenance department, but some of it requires special certification or licensing, like with the bearings or treating the support beams. Big changes to software are often driven by improperly incentivized consultants and their advice is only slightly better than a best guess.
But there's a whole host of things that make bridge building completely different. Bridges rarely get their design changed after they've been installed as doing so is basically a start over except for some of the foundation. Bridges are rarely seen as a vector for compromising the security of everyone in a town just because a bad guy takes an interest and doing so must be done one bridge at a time. A bridge usually decays gracefully leaving cracks and holes, but is still usable whereas software can break in an instant when one of a hundred assumptions is no longer true. Iterative processes for software development are a proven way to get a project much closer to the moving landing point on the opposite shore, which is impossible with bridges which must be a one-and-done by necessity.
All is not lost on the bridge-software metaphor as they both will always have one thing in common: the first thing the accountants cut is maintenance and it takes an embarrassing emergency to get it back.