I was just listening to a podcast from January in which there was much discussion of science being only the best explanation we have now. In fact, being a process rather than a pattern of knowledge.
I am an engineer as you see below. I believe that engineering is different from science, but relies upon it. Some people trained as scientists work as engineers and some universities used to teach what they called Engineering Science, though very few do now.
This microcosm of society provides an interesting study in the establishment of fields of "knowledge" that turns out not to be. Engineers often have massive emotional (and economic) commitment to their established knowledge and so resistance to admitting of changed views. I know that the same is true in science but there is a great community of relatively low level engineers who understand that if their "knowledge" proves to be wrong, peoples lives may be at risk.
My particlular interest is in masonry bridges. One might think we understood them thoroughly after all this time but there is one of the differences between engineering and science. Sometimes you cannot know and therefore have to build systems of protection against the lack of knowledge. Sometimes, though, the knowledge becomes deeply ingrained and even written down in codes of practice which are deemed to be correct. If your design can be shown to satisfy the code you cannot be blamed for failure.
But what if the code is wrong?
Through the enlightenment, the division between engineering and science shifted. People made machines and structures before that but they did it based on established practice and sometimes courage then trial and error. Some of the work of Peter Parler, the imperial mason, in Prague shows spectacular invention but it wasn't limited to him by any means.
Robert Hooke and Christopher Wren were polymath scientists. Both practised as architects. Indeed they shared a practice as architects as well as working independently. In their day, architects were structural engineers. As scientists they had to defend their engineering and it took them considerable effort to discover a way to understand arches. In particular t calculate the abutment needed, or buttress if you will, in the new churches they were building.
Around 1900, Castigliano, a railway engineer, invented a whole new "science" of structural behaviour delivering a new analysis of arches. But railway loads were still small compared with the weight of the bridges and in any case bridges aren't just arches.
Things came to something of a head with the introduction of tanks in the first world war. A general couldn't afford to see maybe 5 or 6 tanks get across a bridge and it then fail, leaving his strength isolated and readily destroyed. The RE couldn't help immediately but over maybe 15 years, work was done to turn Castigliano's efforts into something useable in anger.
They recognised that bridges were complex and did some real tests to calibrate things. They found, but didn't recognise, that the behaviour is inescapably three dimensional and proceeded to create a two dimesnionalising model of an effective strip. That effective strip model is demonstrably wrong but how wrong only became obvious when railway axles reached 25 tonnes and damage started to appear where the assessment tool (the code) said it shouldn't.
The trouble is that engineers still cling to the code. One valid comment is that until we have something better we can’t let go of our only support. But something better needs paying for and if money is not made available for that where do we get to?
Bridges are failing. The engineers in the field map the failure onto their codes and treat with established solutions which are unlikely to work in the longer term. Maybe nothing will happen till a bridge actually falls down under a train.