Interesting post….. I had a small quibble about the Occam part and then get to the more meaty aspects I guess.

Occam’s Razor is explaining unknown phenomenon using the simplest theory; that’s different than engineering the simplest system possible (the KISS/YAGNI principle).

Your last sentence says “Add complexity…only if needed”.

I would disagree here.

The Turing Complex theorem states that any more complex program running on some more complex hardware could be written in a simpler language using a simpler machine, but it would take more time to write and run.

So there is never a “need” — per se — to do the more complex thing — there is just the desire/business case to do so.

There is never a “need” to use separate name spaces, to split up the project, to use a more advanced language, to use a SQL database, all those things *could* be done using a simpler system.

At what point is the simpler system no longer the best solution? That is a judgement call, in some ways, and a business analysis decision on the other hand.

I think people should use the more complex method “when it makes sense” — not just “only if needed”.

Judgement is a large part of the sw dev game, and a big reason why experienced developers bring more to the table.

Best regards,
Jordan