If Flash were standing on the same RFC-based ground that TCP is (Jon Postel was speaking of a principle that implementations of TCP should do) then Dave might have a point.

Heck, I'll be generous and say that Dave might have a point if he were referring to an independent implementation of Flash, and what it ought do.

But mobile Safari is not a flash runtime, it's a web browser, and so its duty to the robustness principle is to be strict about what it sends (HTTP GETs and POSTs) and permissive about what it receives (HTTP responses, HTML/CSS/Javascript documents). It is no more obligated to include a Flash plugin than it is to have a plugin for Iterated Fractal Systems image compressor.

I think it's worth noting that there's an important class of software that needs to be strict about what they receive — compilers — and I often find myself wishing that browsers were more like compilers. But that's just when I get sour thinking about the excessive permissiveness of IE.