Thread · 2 tweets · 17 May 2019

It's more complicated than that. At some point it was described as an SGML schema which does generate a deterministic parser. But it wasn't implemented that way and the structural incentives for a distributed, extensible language don't match that method well (put mildly).

The HTML5 parsing algorithm does indeed map pretty much any arbitrary byte stream into a deterministic tree. As grammars go that may be considered a strange case 😉