I agree with you, however: The WaPo has a veeeeery wide variety of endpoint devices. If they stuck to HTML + CSS it would mean an imperfect output on things like kiosk panels, public news stand displays, ebook readers etc. or needing to maintain multiple output services server side, with potentially different authoring considerations.

With their current approach, they can share one content authoring environment over an API that splits all render concerns to the consuming client.

Sounds a lot like what HTML was supposed to do in the first place doesn’t it… but not all devices support arbitrary html web browsing, nor would you want to serve a full page to most , so the next best option (according to some) is API plus javascript client for the web side of things.

While I don’t think this path is correct, given the tools we have I do see why it is used so frequently.