quest count for the modified version

[tiptoptom]

Use case

It would be nice, if quests made in the modified version could be counted like quests made with the normal version count for the modified version

[Helium314]

This will not happen. Statistics are synchronized with https://github.com/streetcomplete/sc-statistics-service, and this only counts quest that are solved with the original StreetComplete name.
This app is intentionally not using the orignal name, because the changeset comment looks like vanilla StreetComplete was used. And I don't want problematic edits done with this version to be attributed to vanilla StreetComplete.

What could be done:

• run a modified statistics service for this version
• completely remove the synchronization and have everything offline, and hope everything works without synchronization

Both are definitely too much work for me, and none would allow proper comparison with vanilla SC users.

[tiptoptom]

Would it be possible to upload all unmodified quests with the original StreetComplete name to osm and only the modified quests as StreetComplete_modified?

[Helium314]

That would possible, tough it might not be simple to implement.
But it's more theoretical anyway, since all quests allow you to modify all tags via an other answer. And I suspect using different changesets for the same quest but based on the answer likely requires even more work.

[mnalis]

Ah, I wasn't even aware that H3 fork does not count in stats (as I did bulk of my uploads in my fork).
Gamification is still a motivator for me, so this should be addressed before I make most of my commits via it. (and I already switch between OsmAnd in SC all the time, having to switch between 3 apps would be more annoying).

So I see two distinct issues here:

• being able to technically distinguish changes made in upstream vs. changes made in fork.
  This works with current Streetcomplete_modified 43.2 tag, but could work as well with StreetComplete 43.2_ee tag - which would (presumably?) work just fine with counting scores. It would also allow to see how many users are using each version. So I would implement it like that for all quests.
  However, it would not address other issue:

• being fair, so users don't use fork capabilities just so they could gain higher score by cheating.
  This issue might also be problematic, while hopefully much less so then above issue (after all, if user is determined to gain score without regard for data correctness, they might just solve random quests with random answers even in upstream SC. Hopefully gamification of SC is not so addictive).
  Changing name to StreetComplete_modified as @tiptoptom suggests above only for modified quests would solve that issue (in addition for changing version to 43.2_ee for all quests), but it would need some time and effort for investigation and implementation.
  Another (perhaps easier, albeit less granular) option might be to change StreetComplete:quest_type for modified quests instead (e.g. AddPlaygroundAccess_ee instead of AddPlaygroundAccess) if statistics would not count those nonexistent quests (one would need to be checked first, perhaps stats count everything and not only "known" values)?

Also, version seems to be buggy anyway currently (#374) so fix for that could incorporate the first workaround at least (second could be applied later, if it is not deemed huge problem at the moment. Perhaps someone with free time on their hands at that time could use https://osmcha.org/changesets/ to find if changes made with H3 fork seem legit or not, i.e. how important is that second issue)

[Helium314]

Meanwhile I noticed that actually having the created_by name and <app>:quest_type tag different, and thus having statistics counted, is quite simple, as done in https://github.com/matkoniecz/Zazolc/blob/3d18d0f2393429644342420699f164f5aa1a73b7/app/src/main/java/de/westnordost/streetcomplete/ApplicationConstants.kt#L6-L10
So having statistics without claiming to be StreetComplete can be done, even including SCEE-only quests (see streetcomplete#1894).

There are still some things to consider here:

1. Are SCEE capabilities are considered an unfair advantage when comparing star count between SCEE and StreetComplete?
   This is likely okay, as with overlays even in SC the user is not limited to some quests, but can also get points from e.g. micromapping surfaces at intersections, or adjusting shop positions (and reducing resurvey interval also helps for getting more points).
2. Will there be users switching from StreetComplete to SCEE just for the possibility of getting more points?
   Despite my conclusion above, this might be an issue, as additional quests (and adjustable element filters for most quests) allow for more convenient / faster increase of star count.

And in the end it depends on @westnordost whether it's ok to use StreetComplete:quest_type

[westnordost]

Fine for me El 19 de mayo de 2023 10:41:21 CEST, Helium314 ***@***.***> escribió:

Meanwhile I noticed that actually having the `created_by` name and `<app>:quest_type` tag different, and thus having statistics counted, is quite simple, as done in https://github.com/matkoniecz/Zazolc/blob/3d18d0f2393429644342420699f164f5aa1a73b7/app/src/main/java/de/westnordost/streetcomplete/ApplicationConstants.kt#L6-L10 So having statistics without claiming to be StreetComplete can be done, even including SCEE-only quests (see streetcomplete#1894). There are still some things to consider here: 1. Are SCEE capabilities are considered an unfair advantage when comparing star count between SCEE and StreetComplete? This is likely okay, as with overlays even in SC the user is not limited to some quests, but can also get points from e.g. micromapping surfaces at intersections, or adjusting shop positions (and reducing resurvey interval also helps for getting more points). 2. Will there be users switching from StreetComplete to SCEE just for the possibility of getting more points? Despite my conclusion above, this might be an issue, as additional quests (and adjustable element filters for most quests) allow for more convenient / faster increase of star count. And in the end it depends on @westnordost whether it's ok to use `StreetComplete:quest_type` -- Reply to this email directly or view it on GitHub: #370 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

-- Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.

[mnalis]

So having statistics without claiming to be StreetComplete can be done, even including SCEE-only quests (see streetcomplete#1894).

That sounds great; so if I read that correctly upstream StreetComplete will only count quests that it knows about, and SCEE will count a few more for its additional quests? That is even better than expected, I would've been quite happy even with just a first part!

There are still some things to consider here:
Are SCEE capabilities are considered an unfair advantage when comparing star count between SCEE and StreetComplete?

As the majority of the quests is the same, I don't think it should be a big issue. The SCEE-only quests are not counted anyway as far as I understand, so main issue would be people modifying filters so existing quests are asked on more elements just so they can get more stars.However, people who are willing to needlessly spam that way, will likely be perfectly happy to spam in overlay mode in regular SC too.

And people wanting to "cheat" can do it easily by selecting in preferences to "ask quests more often" (but the joke is on them, that is exactly what we'd like them to do anyway 😄)

But I hope SC gamification factor is not so high yet to provoke damaging / needlessly spamming the OSM; after all there are free games just a click away which are much more addictive if one does not really care about improving OSM data as well.

Will there be users switching from StreetComplete to SCEE just for the possibility of getting more points?

Well, I'd hope it would not be an issue:

• SCEE is not available on Google Play, so it means that most people won't even know/consider it
• of those who do know about / consider SCEE, some will still prefer simplicity/popularity/safety of the original.
• most eople who use f-droid are also more likely to care about open projects (hopefully to the point where they care more about those open projects than about the SC score)
• as noted above, I don't think SC score is so addictive that people will consider changing apps in order to slightly increase their score. I mean, we have like 5 👍 on this issue which would allow people to cheat, so even if 100% of us did so just so we can jump ship to bump up our SC score unfairly (which, needless to say, I find highly doubtful! 😃), it would still not be the end of the world.

[Helium314]

How important is having the star count on the main screen for you?
I'm thinking about keeping it hidden to have some clear difference to StreetComplete that is always visible in screenshots.

But I seem to care less about statistics / star count than most others, so I shouldn't decide all by myself.

[tiptoptom]

Do we want to vote with reactions?
👍 the star counter on the main screen should remain
😕 I do not care
👎 the star counter should no longer be visible on the main screen

[heragnar]

As a user of both apps, I believe that the inclusion of a star counter or other popularity contest has no place in an app designated as the Expert Edition. Right now it only clutters the interface for advanced or expert users, who make up the majority of the user base, and can cause confusion for regular users who may expect certain functionality based on the presence of the star counter. This feature does not contribute to the overall statistics of upstream StreetComplete anyway. Therefore, removing the star counter would improve the user experience for all users, including advanced and regular users, by eliminating the potential for confusion and distractions.

[mnalis]

As a user of both apps, I believe that the inclusion of a star counter or other popularity contest has no place in an app designated as the Expert Edition

This I disagree with. Just because SCEE allows the user to tweak UI in faster-to-use-but-more-confusing-way and some features might expect some OSM knowledge about tags, does not mean that gamification suddenly stops being a thing.

Right now it only clutters the interface for advanced or expert users, who make up the majority of the user base,

This I can agree with somewhat, which is why I am not strongly opposed to removing the star from main screen (but still absolutely want the starcount and other statistics to continue being available in My profile, and edits made with SCEE to count in both upstream SC and SCEE!)

and can cause confusion for regular users who may expect certain functionality based on the presence of the star counter.

Hmmm, what do you mean there exactly? What functionality is based on the presence of the star counter?

This feature does not contribute to the overall statistics of upstream StreetComplete anyway

I'm not quite sure what you mean there, but if mean to imply that editing with SCEE does not count toward upstream StreetComplete "score", that should no longer be the case (i.e. SCEE edits nowadays should get counted in both SCEE and upstream StreetComplete, see #370 (comment))