[ellen364]

That seems like the best case scenario — the over-engineering is costly upfront and worth it later on. But if there’s a problem in the over-engineering, e.g. mistaken assumptions, fixing it is another round of expense.

So over-engineering is “expensive now and hopefully cheap later”, while under-engineering is “cheap now and maybe expensive later”. In one scenario you get 1-2 rounds of expensive. In the other you only get 0-1 rounds of expensive.

With that in mind, under-engineering the first implementation might be the sensible default choice.