[ochoseis]

One benefit software has over physical systems is that you can easily document it right inside the code with comments and types to make the intention clearer. This isn’t foolproof, especially in dynamic languages like Python, but helps a lot.

The analogy for the load bearing stud might be a hackathon project that never expected to see production. In reality, a lot of what we do is hack on something until it barely works, and move on to the next thing.

[MilStdJunkie]

Yuuuuuup. This precise set of problems is why Systems Engineering was adopted in Aerospace and Defense, because maintenance plans need to know what the "loads" are for each replaceable (spareable) unit or assembly.

Today's SE has wandered far, far afield from original goals, tragically enough, but that was the original conception. One of the reasons for today's relatively toothless SE departments is the rise of finance into maintenance planning. Inventory depreciation is a cruel mistress, and "what gets spared" is rarely a SE judgement these days, at least in my experience. This has predictable results, but is partially offset by the exceptionally high bar for aerospace maintenance staff, who are generally pretty damn badass compared to, say, a washing machine repairman. Finance, naturally, would like to knock that bar down a few pegs, too.

[eschneider]

That's easy to do, if it's _intentional_. I'm always amazed at how often I see systems rate limited by some seemingly "decorative" upstream component, that when removed, causes everything else to run amok. :/