Expose an API to be able to get all app starting points (origin + launch_path) by device optimization

RESOLVED INVALID

Status

--
enhancement
RESOLVED INVALID
6 years ago
6 years ago

People

(Reporter: jsmith, Unassigned)

Tracking

Points:
---

Details

(Reporter)

Description

6 years ago
For mobile web compatibility testing for gecko, a really helpful need from marketplace is the ability to query marketplace to get all app starting points on launch (origin + launch_path). With this exposed API, we would be able to let the ateam's tool or Aaron Train's screenshot tool to be able to get these URLs for app starting points and generate useful data for this specific tools, which includes:

- User Agent Sniffing rates per mobile app listed on marketplace
- Screenshots for apps to analyze web compatibility quality across multiple browsers
- Layout compatibility analysis to capture webkit aliases problems
- And other compatibility analysis use cases as needed
(Reporter)

Updated

6 years ago
Severity: normal → enhancement
(Reporter)

Comment 1

6 years ago
Although I know this probably wouldn't block basecamp, this is strong "want," as it could provide very useful data we could use for web compatibility testing of gecko across major platforms.
I don't think that we're going to get much useful compatibility data from the Mozilla marketplace. Marketplace apps should have gone through some sort of QA testing (even if just simply testing by the dev) and, at this point, necessarily run on Gecko. The developers of these apps are already engaged with Mozilla technology. We also have a marketplace review process in place to ensure that apps meet some minimum quality standard. Am I missing something?
(Reporter)

Comment 3

6 years ago
(In reply to Lawrence Mandel [:lmandel] from comment #2)
> I don't think that we're going to get much useful compatibility data from
> the Mozilla marketplace. Marketplace apps should have gone through some sort
> of QA testing (even if just simply testing by the dev) and, at this point,
> necessarily run on Gecko. The developers of these apps are already engaged
> with Mozilla technology. We also have a marketplace review process in place
> to ensure that apps meet some minimum quality standard. Am I missing
> something?

Not every single app that lists itself as mobile on marketplace happens to work well on mobile. Most early reviews were heavily done on desktop, not on a mobile device I believe (that might have changed more recently, but ask Andrew about it), so we might have apps in this situation. I already know one example, although it's not a web compat issue (Sinuous, which just so happens to not look good on mobile at all). 

Wouldn't the marketplace review process find the data we collect from our two respective tools useful here though? Perhaps Andrew could comment here on his opinion on web compat statistics for marketplace app reviews.
(In reply to Jason Smith [:jsmith] from comment #3)
> Wouldn't the marketplace review process find the data we collect from our
> two respective tools useful here though? Perhaps Andrew could comment here
> on his opinion on web compat statistics for marketplace app reviews.

I think this is a good point. The marketplace review team may find it helpful to get an automated report about app submissions. And, new apps definitely should be tested on mobile as our initial launch of the marketplace is with Firefox OS.
(In reply to Jason Smith [:jsmith] from comment #3)
> (In reply to Lawrence Mandel [:lmandel] from comment #2)
> > I don't think that we're going to get much useful compatibility data from
> > the Mozilla marketplace. Marketplace apps should have gone through some sort
> > of QA testing (even if just simply testing by the dev) and, at this point,
> > necessarily run on Gecko. The developers of these apps are already engaged
> > with Mozilla technology. We also have a marketplace review process in place
> > to ensure that apps meet some minimum quality standard. Am I missing
> > something?
> 
> Not every single app that lists itself as mobile on marketplace happens to
> work well on mobile. Most early reviews were heavily done on desktop, not on
> a mobile device I believe (that might have changed more recently, but ask
> Andrew about it), so we might have apps in this situation.

Yeah, all the early reviews were just on Desktop.  Recently testing has been done on Android also.  - Occasionally I've not tested on Android if the app was fairly simple, and designed well enough (large targets to press, etc) and seemed like it would work the same as Desktop. 

I've rejected Apps for not working properly on Mobile a few times (or suggested they remove Mobile support).  The main culprit is games that rely on keys or mouse pointer position.  Unfortunately, in general, crappy design (that still just about works) isn't one of the criteria we use for rejection.

> I already know
> one example, although it's not a web compat issue (Sinuous, which just so
> happens to not look good on mobile at all). 
> 
> Wouldn't the marketplace review process find the data we collect from our
> two respective tools useful here though? Perhaps Andrew could comment here
> on his opinion on web compat statistics for marketplace app reviews.

Any data would be of interest though I'm not sure exactly what action we would/could take - what kind of scenario are you imagining?
(Reporter)

Comment 6

6 years ago
(In reply to Andrew Williamson [:eviljeff] from comment #5)
> (In reply to Jason Smith [:jsmith] from comment #3)
> > (In reply to Lawrence Mandel [:lmandel] from comment #2)
> > > I don't think that we're going to get much useful compatibility data from
> > > the Mozilla marketplace. Marketplace apps should have gone through some sort
> > > of QA testing (even if just simply testing by the dev) and, at this point,
> > > necessarily run on Gecko. The developers of these apps are already engaged
> > > with Mozilla technology. We also have a marketplace review process in place
> > > to ensure that apps meet some minimum quality standard. Am I missing
> > > something?
> > 
> > Not every single app that lists itself as mobile on marketplace happens to
> > work well on mobile. Most early reviews were heavily done on desktop, not on
> > a mobile device I believe (that might have changed more recently, but ask
> > Andrew about it), so we might have apps in this situation.
> 
> Yeah, all the early reviews were just on Desktop.  Recently testing has been
> done on Android also.  - Occasionally I've not tested on Android if the app
> was fairly simple, and designed well enough (large targets to press, etc)
> and seemed like it would work the same as Desktop. 
> 
> I've rejected Apps for not working properly on Mobile a few times (or
> suggested they remove Mobile support).  The main culprit is games that rely
> on keys or mouse pointer position.  Unfortunately, in general, crappy design
> (that still just about works) isn't one of the criteria we use for rejection.
> 
> > I already know
> > one example, although it's not a web compat issue (Sinuous, which just so
> > happens to not look good on mobile at all). 
> > 
> > Wouldn't the marketplace review process find the data we collect from our
> > two respective tools useful here though? Perhaps Andrew could comment here
> > on his opinion on web compat statistics for marketplace app reviews.
> 
> Any data would be of interest though I'm not sure exactly what action we
> would/could take - what kind of scenario are you imagining?

Hmm okay (so crappy design does not affect it), so maybe this won't affect the review process. My original thought was having a way to know if we have "good" quality apps that look good on gecko being the rationale for having this through the mobile web compat analysis, but I'm rethinking now that maybe, as this won't affect the review process.

I'll resolve this as invalid for now, but if I get a different idea, I'll file a bug again.
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → INVALID
It would still be good to know how many 'good' mobile apps we have I suppose and many app developers do take feedback on board if we point out an issue to them.  The data could still be used to encourage developers to do better if you want to re-target the audience to them rather than us.
You need to log in before you can comment on or make changes to this bug.