|
|
|
|
|
by sour-taste
2523 days ago
|
|
I 100% agree with this article, but I suspect it will not be well received here. HN is so focused on innovation at all costs that anything that goes against that will be rejected. I think that one of the issues with doing as the author suggests and thinking of technology as a transformation is that it can be hard to come up with the downsides of your new tech. Engineers will constantly be encouraged by management and customers to deliver a solution, and often saying 'this is my solution, but it comes with caveats' will be frowned upon. Additionally, it's clear in retrospect that the expansion of the city the grand canal allowed would lead to depletion of the ground water which would lead to sinking of the city and the canal being inneffective, but was that known at the time? Could engineers working on a canal project have anticipated socioeconomic trends like this? I'm not saying they couldn't have, just that it's difficult. |
|
Mexico City has grown from 4 million in 1951 (the time of flood mentioned in the first paragraph) to over 20 million today. Few civil engineering projects solve problems 70 years in the future for a city five times as large. As far as I can tell from the article, the engineers did anticipate population growth and designed a system that would last for decades; furthermore, they monitored the system, were aware of stresses and possible failure points, and took several appropriate corrective actions as the decades went on. All of this is just from reading the article.
If at my job I criticized a design by saying, "this design is terrible because if it is wildly successful and gives us 5X growth, then 70 years from now it might require some rework" I would be laughed at. If I followed it up by suggesting that this reflected some sort of fundamental problem with the methods of engineering I would probably not be taken seriously. Engineers are not godlike miracle workers, setting in motion an ineffable plan that somehow make the world holistically better over an indefinite time frame. It's OK to just solve the tractable problems in front of us and to get a good couple of decades out of a system.