1245203 - consider warning users that Android web runtime will be disabled

Status: RESOLVED WONTFIX

(Firefox for Android Graveyard :: Web Apps (PWAs), defect)

Description

[Myk Melez [:myk] [@mykmelez]]

It might be worth warning users that the Android web runtime will be disabled, and giving them a link to a document that describes the change in more detail and suggests alternatives, before we actually disable the runtime, to mitigate the impact of the change.

For example, we might display a warning dialog with a link to a page when the user runs an Open Web App in some release of the browser, then remove the runtime in the next release, so users see the warning when they upgrade Firefox to the first release and can follow the link for more information before they upgrade Firefox to the next release.

Comment 1

[:Margaret Leibovic]

Similar to the notifications we've been showing when we're dropping support for a version of Android, we could also show a system notification when the user launches a web app. But if we're worried this wouldn't be noticeable enough, maybe a dialog would be a better idea.

We could ship this notification in 47, then follow up with the changes to actually drop support in 48.

Comment 2

[Anthony Lam (:antlam) Find me, @antlam on Slack!]

Hm, I'm not sure this is on the same level as dropping support for a version of Android. As a user, never getting updates for Firefox on my phone is concerning. But I also understand what that means because my other apps update and I'm familiar with the value that an update brings.

Just to play devils advocate here, do we think our users would feel the same way about this? A warning is great but if there's nothing they can do themselves, it feels like overkill a bit. Especially because as I understand it, this is something site authors have control over, not our users.

I wonder what the change most users will notice is... and do we expect this to be something most users will understand or need to? Could you point me to a common site that will be majorly affected by this change? that might help me frame this a bit more. Thanks!

Comment 3

[Myk Melez [:myk] [@mykmelez]]

(In reply to :Margaret Leibovic from comment #1)
> Similar to the notifications we've been showing when we're dropping support
> for a version of Android, we could also show a system notification when the
> user launches a web app. But if we're worried this wouldn't be noticeable
> enough, maybe a dialog would be a better idea.

Erm, a system notification would probably be plenty noticeable.  My idea about a dialog mostly betrays a desktop-centric thinking (see bug 1245204 for the virtually identical desktop equivalent).  Although a dialog does enable us to include a checkbox like "[✓] Show this warning the next time", if that's something we want.


> We could ship this notification in 47, then follow up with the changes to
> actually drop support in 48.

Indeed. That'd give users about six weeks advance notice of the change, which they can use to switch to in-browser or native versions of the apps (where available) or alternative apps.

Comment 4

[Myk Melez [:myk] [@mykmelez]]

(In reply to Anthony Lam (:antlam) from comment #2)
> Hm, I'm not sure this is on the same level as dropping support for a version
> of Android. As a user, never getting updates for Firefox on my phone is
> concerning. But I also understand what that means because my other apps
> update and I'm familiar with the value that an update brings.

Note that in this case the apps won't merely stop updating; they'll stop working entirely.  After the user updates to a version of Firefox that doesn't support Open Web Apps, launching an Open Web App will fail.


> Just to play devils advocate here, do we think our users would feel the same
> way about this?

Note that this would only appear for users who have Open Web Apps, and it would only appear (at least as I've been thinking about it, although I'm open to alternatives) when a user actually launches one of their apps.  So most Fennec users would never see this.  It would only impose on users who actually have something to lose, because they've installed apps, and they use them.

And given that context, yes, I do think those users would want to know that apps they actually use are going to stop working suddenly at some point in the future. ;-)


> A warning is great but if there's nothing they can do
> themselves, it feels like overkill a bit. Especially because as I understand
> it, this is something site authors have control over, not our users.

App developers have control over updates, so they can also break users.  But in this case, we're the ones breaking them.  And there is something that users can do about it: they can switch to an in-browser version of the app, switch to a native version of the app, or switch to an alternative app.

Of course they can also do all these things *after* an app stops working.  But it would be a less abrupt transition to warn them in advance that the app will stop working, and give them a period of time to make the transition, during which time the app will continue to work; rather than for their apps to stop working, at which point they have to make the transition immediately to continue having whatever functionality the app provides, regardless of whether they're ready to do so.


> I wonder what the change most users will notice is... and do we expect this
> to be something most users will understand or need to? Could you point me to
> a common site that will be majorly affected by this change? that might help
> me frame this a bit more. Thanks!

Unfortunately, I don't have great stats on the number of users, nor the most common apps in use, nor the number of active installations.  I don't expect usage to be high, but even if it's low, the impact on those folks who are using apps may be significant.

That being said, I'm not convinced that we *should* do this, which is why I filed the bug to "consider" it rather than to do it!  It isn't that I don't think it's worthwhile.  Mostly I'm just concerned about the engineering trade-off of doing this versus something else.

Nevertheless, I do want to hear countervailing points of view.  So please do play the devil's advocate, or even just disagree, if that's what you think!

Comment 5

[:Margaret Leibovic]

We do have telemetry that tells us when Firefox was launched via a web app, as opposed to tapping on the Firefox icon, or opening via a URL:
http://mzl.la/1KX6uVG

Data says people basically don't use this feature (web app launches are 0.09% of all app launches).

There may be reasons to mistrust this data - it's not clear that we're receiving all telemetry uploads, since these pings do come from separate web app profiles. However, this is still such a small percentage, even if it's off by an order of magnitude.

It could be jarring to have your web app stop working without any waning, if you are one of these users, but we also stopped promoting the marketplace or about:apps long ago, so this shouldn't surprise people (i.e. they didn't *just* install one of these apps). App retention is crazy low on mobile, so I'm be surprised if people are still using these apps much at all, given "installs" should have stopped a long time ago.

Thinking about this more, and looking at this data, I think we should just prepare a SUMO article about the transition and be ready to address user concerns as they happen. I'd rather get this removal over with sooner rather than later, and adding a string for this message would be a blocker to starting this work in 47 (or sooner, if we can uplift changes).

Comment 6

[Barbara Bermes [:barbara] - NI please!]

+1 SUMO article should be sufficient considering the impact and the low usage of this.

Comment 7

[Myk Melez [:myk] [@mykmelez]]

(In reply to :Margaret Leibovic from comment #5)
> There may be reasons to mistrust this data - it's not clear that we're
> receiving all telemetry uploads, since these pings do come from separate web
> app profiles. However, this is still such a small percentage, even if it's
> off by an order of magnitude.

Indeed, this data doesn't surprise me.  Even if it's incomplete, app usage is still likely to be low.


> It could be jarring to have your web app stop working without any waning, if
> you are one of these users, but we also stopped promoting the marketplace or
> about:apps long ago, so this shouldn't surprise people (i.e. they didn't
> *just* install one of these apps). App retention is crazy low on mobile, so
> I'm be surprised if people are still using these apps much at all, given
> "installs" should have stopped a long time ago.

Indeed, although users can still install apps from Marketplace (or any other site, for apps that don't request privileged APIs).  So I don't think installs have entirely stopped.


> Thinking about this more, and looking at this data, I think we should just
> prepare a SUMO article about the transition and be ready to address user
> concerns as they happen. I'd rather get this removal over with sooner rather
> than later, and adding a string for this message would be a blocker to
> starting this work in 47 (or sooner, if we can uplift changes).

Ok, sounds good, then I'll wontfix this bug.