On 8/8/18 12:53 PM, Boris Zbarsky wrote:
On 8/8/18 12:43 PM, Chris Pearce wrote:
On Tuesday, August 7, 2018 at 7:51:50 PM UTC-7, Jan-Ivar Bruaroey wrote:
To clarify, I care about Netflix, which is why I question giving up on
persisting autoplay for them, which is what allowedToPlay does.
So I have a question. Would it be at all useful to have an explicit way
for a site to ask for a persistent permission grant (as opposed to the
implicit "just try to play something")? Would netflix use that?
Generally, our position [1] is that standardized permission-oriented
APIs are bad for the web and Firefox. Because:
1) It would encourage permission bundling, e.g. the dreaded "Approve all
56 permissions to use this Android Flashlight app", of olden days. It is
generally agreed these days that asking as late as possible ensures the
most context for users to know what specific features they'll miss out
on if they decline. The way we do that is embed prompts in the access APIs.
2) It would mean effectively standardize one permission model for the
web (probably Chrome's). We consider this a bad thing, because
permission is still an area of differentiation and innovation between
browsers. To the extent we attempt this WITHOUT standardizing, we end up
exposing non-standard user agent behavior, and then we always lose,
because many web developers will then test against the dominant
browser's permission model only, and things end up working poorly in
Firefox, unless we mimic Chrome's model exactly.
I also think it's premature to say that mapping block autoplay to a
permission model is even a perfect fit, so I wouldn't draw any general
lessons about permission from block autoplay. That would be the tail
wagging the dog IMHO.
Or would they avoid it for the same reasons as their current avoidance of
the prompt?
I'm missing why this would make the prompt any more palatable to sites
that "vehemently" don't want it. I'm also not sure how we would know,
since this is effectively a negotiation. E.g. if you asked me, and I
didn't want it, I'd say "yes" and then never call it, to get the prompt
out of the way.
There's a hard problem here at heart where the APIs are being designed
around the UX, and the different browsers have very different UX goals,
afaict.
Again I wouldn't generalize from this case. "The APIs" are designed
around requesting resources, not permissions. User agents are separate
from the web, and there's value in allowing diversity among user agents
for hard problems, like guiding users with different privacy needs, ad
blocking, permissions, hard problems like that which are not inherently
rooted in how web content is shaped, but more how users have different
preferences for consuming (viewing and interacting with) said content.
-Boris
[1]
https://github.com/mozilla/standards-positions/issues/19#issuecomment-370039662
_______________________________________________
dev-platform mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-platform
