Re: [w3c/permissions] Drop .request() (#83)

Coming in to this after a break (hello from Almaty!).

The problem I have with this discussion so far is the presumption (in various posts) that a particular model is somehow wrong, or that this is something where a compromise is necessary.

I see us wasting a lot of time on this, and not actually making any progress.  That is because we disagree.  I was given the impression that variations in how browsers interact with users was an important property to preserve.

That means that disagreement isn't a real problem.  Well, it is as long as `.request()` continues to exist because I think we've proven that it can't be implemented consistently without forcing changes on browsers that they don't want (granted, it's clear that Mozilla aren't in agreement on this point, though that's not news).

I don't find the polyfill convincing either way here.  It merely takes the current APIs and switches direct effects with secondary effects.

So let's talk about effects on implementations: say that the purpose of `.request()` is to produce whatever UX will turn a call to `.query()` from "prompt" to "granted" (assuming that they are made within the same scope, and that the user doesn't change their mind, blah, blah, ...).   In implementations that have a "remember this choice" option (as Firefox does): are they forced to fix that toggle in the "remember" position?

I haven't seen a satisfactory answer to that; as I see it, this would force Firefox to do that, which I don't think we're prepared to do.  Also, implementations that present a choice (as with `getUserMedia()`), but associate persistent permission with all devices (again, Firefox), are they obligated to remove the choice so that the intent of the prompt is properly clear?

It's clear that if there is no possibility of persisting permissions, then `.request()` will always fail, but if a persistent grant is ever conditional, this API produces some surprising effects.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/w3c/permissions/issues/83#issuecomment-243978970