On 5/28/15, 10:08 AM, "Frédéric THOMAS" <webdoubl...@hotmail.com> wrote:
>Btw, I would like to come back on this: > >> 1) figure out when to load the polyfill > >For ES6: All the time at the moment > >> 2) worry about bugs in the polyfill > >I guess like any other thirdparty we are using and those polyfills we are >talking about are open source, it means as well there are people to >maintain and fix them until ES6 is supported by the browser. > >> 3) have different debug experiences in different browsers > >I'm mot sure that's a problem, if a dev needs to debug at this level, >they should rely on 2, othewise they would debug in ES6, which is more >comfortable ? > >> 4) bundle the polyfills in the release > >Not sure, I guess they could be loaded directly from their github >location. > >> 5) manage the licenses and other documentation around the polyfills. > >Indeed and I'm not good with that. > >Did I miss something ? Nope, those sound like the right answer, but it sounds like more work than just staying with ES5. So ES5 would be my preference, but I’ll go with the majority. Regarding the debug experience: the more runtime configurations you have to support, the more time you will spend investigating those bugs where it works for you but not for the bug filer. Odds are you’ll end up stepping into the polyfill to see if it is the reason. Again, if there is enough pay-off for moving to ES6 great, but otherwise, that’s just more work for us. -Alex
