Thanks! This is covered at the end of the post, but I can elaborate a bit here. HTM mirrors JSX syntax, which means it requires closing tags for every element.

Thanks! This is covered at the end of the post, but I can elaborate a bit here. HTM mirrors JSX syntax, which means it requires closing tags for every element. We wanted HTML5 syntax instead, which means our parser needs to handle void elements and missing close tags, and a variety of other edge cases (such as “bogus” comments) whose behavior is defined by the HTML5 specification. Another important feature for us is contextual namespaces. If you have a G element within HTML, it’s an unknown element, but within SVG it’s a group element. HTM (and JSX, HyperScript, and similar approaches) build the DOM from leaves up, so at the time you create the element, you don’t know what the parent is and thus can’t infer the appropriate namespace. That’s fixable by changing what HTM passes to the “pragma”, but it would require changing the signature.