Semantic Compression

I'd be a lot more convinced if he'd posted a procedural version of the data model whose OOP implementation he mocked, rather than the simplest UI code he could find. Or if the UI example hadn't centered around a "struct with member functions", which is oh so very very completely totally very different from an object.

I'll grant that OOP is often very poorly taught in schools, and very poorly written about in blog posts with contrived examples. But insisting dogmatically on procedural code to model an object is just as silly as insisting on object-oriented code to program a procedure. There's room in the toolbox for both, with even a little left over for some functional programming once in a while.