I'm not speaking from the viewpoint of the consumer of the browser. I'm speaking from the viewpoint of the CGI application. It emits a set of containers (
PerlMouse is correct) and those tags contain additional data. Those containers and the data they contain are all specified in plaintext and that plaintext is optionally parsed by the browser's rendering engine. The browser doesn't have to be a GUI browser a la IE or FF. It can be something as simple as Notepad or some other editor. That is a legitimate browsing client for a given HTTP stream.
As for HTML being a programming language, I think Stroustrop is trying to find ducks that dig. It just doesn't pass the sniff test.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?