Sure, your points on their differences are correct, I agree with that. But I still think my comparison holds.
Basically the comparison was made to convery Sturgeon's Law; that 90% of everything made is crap.
In that sense it is like evolution: A numbers game. While not explicitly random, distributing risk among groups going in different direction could be an efficient way forward, when counting for what would be lost with some central governance. You then select the optimal direction by results.
90% of all bridges and buildings, powerstations, cranes and heavy machines are not crap. Engineering is a responsible job, it's nothing at all like making clay vases or plates where if you throw away 9 out of 10 of the things you make isn't going to cause anybody to suffer other than yourself.
What if they are crap, but just by a metric that isn't "falls apart immediately"?
I watched my state government build an overpass that wasn't needed and didn't even have an _exit_ for 3 years. It required multiple road closures and for several months made traffic awful and my drive unsafe. I call that "crap".
I'm sure that there are organizations that will mess up if you give them enough rope. But that says nothing about the overpass itself and 'doesn't fall apart immediately' is not a standard that any civil engineer will want to be associated with, at least, not in the developed world and likely not in the third world either (but there due to resource constraints and corruption there may be a difference between what should have been done and what was actually done).
Sure, but that's because of safety necessity, not because of optimal solution. If building bridges, powerstations, cranes, etc etc were essentially free outside of drawing/planning them, and could be deployed outside of "production" infrastructure, I'm sure that industry would be lightyears ahead of where it is.
Software is rapidly becoming just as critical, and at some point may be more critical.
The ability to copy a product is a thing that works to your advantage after you've created it, the fact that each bridge and each building is engineered with slowly changing techniques and standards is what makes engineering a solid profession. I think that we probably have had too much change in too short a time thrown at us in the software related professions, if our hardware and other engineered products would perform at the same reliability level and with the same degree of resource consumption that our software routinely gets away with - not even looking at the risks associated with deploying crappy software - we would have returned the product to the manufacturer.
But somehow we've collectively managed to convince our customers that this is the best we can do and that there is no honor in going slower but safer. This is all fine until software becomes critical to our survival, and we're definitely passing some threshold in that respect and as an industry I don't think we are ready to accept that responsibility without making some major changes in how we go about our daily work.
Basically the comparison was made to convery Sturgeon's Law; that 90% of everything made is crap.
In that sense it is like evolution: A numbers game. While not explicitly random, distributing risk among groups going in different direction could be an efficient way forward, when counting for what would be lost with some central governance. You then select the optimal direction by results.