The whole "semantic HTML" debate really seems like the two-state solution of the web world. Browser vendors are profoundly uninterested in the whole topic. It c

The whole "semantic HTML" debate really seems like the two-state solution of the web world. Browser vendors are profoundly uninterested in the whole topic. It could be relevant for search engines in theory - but in practice, the only remaining relevant engines (that happen to be identical to the browser vendors) use proprietary algorithms that deal with messy, un-semantic HTML just fine. (Or, if they have a specific need for semantics, they will just invent their own markup). That leaves a long tail of "specialized" HTML consumers and usecases, such as screen readers, niche search engines, open source projects or scientific projects that actually would benefit from consistent, widely-used semantic HTML (to an extent), but don't have the market power to push the world in that direction. However, those still have the "founding vision" of HTML on their side, that browsers rely on for their brand reputation. So browser vendors can't simply say out loud "look guys, you lost. HTML today is for showing stuff in a browser and running javascript and nothing else, deal with it". The result seems to be that the "semantic" people get to claim the "moral" victory of how HTML "should" be. They are encouraged to spread their vision at countless web Dev conferences, under the enthusiastic support of the browser vendors - while they are doing absolutely zero to support that vision in actual browser implementations.