Page MenuHomePhabricator

Sunset Nearby
Closed, ResolvedPublic1 Estimated Story Points

Description

For numerous reasons - most prominently, low usage not justifying the maintenance overheads - it is time to sunset Nearby.

Event Timeline

@Johan We will need some community engagement assistance on this one.

Charlotte set the point value for this task to 1.Jul 22 2019, 4:12 PM
Charlotte renamed this task from Remove Nearby to Sunset Nearby.Jul 22 2019, 4:56 PM

For community communication outside of the app, this should be fairly easy, something like this:

  • Mentioning it in Tech News.
  • Posting on update in the documentation.

Potentially one would want to mention it in the app as well?

To refresh my memory, I thought about the reasons why we’re removing the “Nearby” feature in the first place. Also, I think it’ll be crucial to communicate this change as transparent and clearly as possible to the community:

  • Our main design and development focus in FY 18/19 is to encourage Android app users to start editing by presenting them with more opportunities to make micro-contributions.
  • To improve “Nearby” in a satisfying way it would take up too much of our resources at this point. The feature in its current state does not live up to our expectation of delivering highly usable products.
  • Bandwidth usage when downloading/updating the app: Nearby, respectively the “Open Street Maps” component by itself takes up XX percentage the app’s size (@Dbrant, could you provide details here? Thanks). As the app’s used a lot in areas where internet connectivity is limited and speed is low, we’ll make downloading/updating more accessible by removing the feature.
  • And last but not least, the usage of “Nearby” has been very low. Only 0.03% of all app users are using the feature. Here’s the stat that Dmitry pulled a while ago. (Could you provide some more details about the time period of this data too @Dbrant? Thanks.)

image.png (425×1 px, 102 KB)

Did I forget something @Charlotte? And thanks for the support here @Johan, it’s very appreciated!

This is one of those times when I think a less elaborate argument would actually be more suitable and realistic. While everything you mention is true, @schoenbaechler, the very, very tiny minority of people who care about Nearby will likely not be satisfied with the argument that updating it will take too many resources away from editing (which they may not use or care about). Since they presently use and enjoy Nearby, they will also presumably not care that we have no plans to improve it in the future. So the final argument - the fact that Nearby only represents 0.03% of pageviews, and it is not worth it for us to maintain - is really the most salient one.

I will update the documentation to reflect the fact that the feature is being sunset, and then we can link it from Tech News.

Low usage is a terrible metric to remove features. Valuable usage is what matters. For instance, from a 'global' perspective most Sysop gadgets are never used.. but removing them would break many workflows for admins which in turn affects everyone.

Personally I think that part of what is wrong with Nearby is that it is not what i care about generally. I hardly ever stand at a spot and think "i wonder what WP has around me" (not even as an editor). Now "where is this item and what is around it, and oh look pretty pictures" is a different story.

Low usage is a terrible metric to remove features. Valuable usage is what matters. For instance, from a 'global' perspective most Sysop gadgets are never used.. but removing them would break many workflows for admins which in turn affects everyone.

In this case we're speaking about an end-user feature that was meant to attract casual browsers, and for such end-user features, usage is an important metric. (Sysop gadgets are another case entirely, and subject to another set of considerations, as you mention.) In the case of Nearby, the very low usage that we see (0.03% of pageviews through the app) is an indication of precisely what you are saying - that the feature is not something that users care about, generally. So I think we are in agreement. :)

ABorbaWMF added a subscriber: ABorbaWMF.

Looks gone to me on phone and tablet on 2.7.50300-alpha-2019-10-19

Potentially one would want to mention it in the app as well?

Yes, absolutely. Or at least something at all. I was just searching for it everywhere—on the bottom navigation, in the hamburger menu… No luck. I found this task by searching through GitHub commits, which is far from the abilities of an average user. (I don’t remember reading about this in the Google Play release notes, either, nor in the Tech News of the last two months. The only info about Nearby in the FAQ is why are some permissions required, nothing about the removal.) I also don’t know about any replacement offered by any Wikimedia service. (Special:Nearby is not a replacement, as it doesn’t show a map, and doesn’t allow searching for any other place than what my browser tells it. It was nice to browse around in the city and find random articles about things out there, which is impossible with Special:Nearby.)

@Charlotte: Updating the FAQ is good, but not enough for a such prominent feature (actually it’s less prominent than Nearby was). I think the best way would be:

  • Creating a replacement, e.g. on Toolforge. This is the most important, no FAQ entry can replace this step. (A web service lacks many of the mentioned disadvantages like app size impact, and the extra data usage is also more obvious.)
  • Deprecating the feature—still providing it, but adding a warning that it will be removed soon, and pointing to the FAQ and/or the replacement tool.
  • Waiting at least some weeks to reach as many users of this feature as possible.
  • Removing it, but still keeping some notice: mentioning the removal in Play Store release notes, adding a home page card, and probably mentioning it in Tech News (these should point to the FAQ and/or the replacement as well).
  • Eventually removing the home page card.

Yes, it takes time, but I think it’s necessary to avoid user confusion like in my case.