[ben336]

You can't use feature detection with a simple script tag. It's calling in a script and the cdn chooses the file to send based on the user-agent. No Javascript is run that would allow it to feature detect before choosing to send the load. There may be feature detection afterwards of course.

[coldnebo]

I think we're saying that most modern polyfills do a feature check before installing the polyfill... thus if you use Chrome with an IE agent, the polyfill still checks before implementing -- it's just a little slower than not installing the polyfill to begin with. Anyway, given the percentage of people who do this, UA detection seems like an acceptable tradeoff.

[ben336]

Agreed, feature detecting afterwards makes sense. Others seem to be suggesting feature detecting before requesting the polyfill script.

[Drakim]

While the roundabout trip would be a waste, it shouldn't be hard to just include a quick and simple feature detection before the polyfill is injected. I don't see any reason why one wouldn't do it.

[ben336]

Quick and simple meaning... are you checking for every feature back to ES3?

The point of this service is that its not requiring the user to know what feature to look for. They'll polyfill anything thats needed. To do feature detection to the level they're trying to serve you'd have to check EVERY API that needs polyfilled on an in-use browser. Full feature detection also involves looking for apis that exist but don't follow the spec. Checking every single API possible on every browser would not be quick and simple. User agents allow them to settle on a list of known needed polyfills without any code needed for the user beyond a simple script tag.

There are definitely tradeoffs to their approach but I can't say "I don't see any reason" why you wouldn't feature detect for this.