> This point of view is based on a fundamental misunderstanding of what the web is. HTML is more concerned with semantic abstractions than the specifics of how a page renders (use a pdf if you want to control the layout). This was an important design choice because it recognizes that the client is responsible for the rendering, which is free to completely ignore the suggestions made by the page.
You are describing the web as it was 25 years ago. No modern browser works this way.
That's the fundamental misunderstanding I was talking about. You don't know what browser is being used. Ever. You only know what the user agent decided to describe itself as and the requests it made to your servers. None of that tells you anything about what happened on the remote computer.
Besides, "modern browser" is a complete undefined term that is obviously subject to a lot of interpretation and opinion. Usually this term is used as either a euphemism for "I really want to put spyware on my website" or "I only use badly-designed tools".
You are describing the web as it was 25 years ago. No modern browser works this way.