When rules are rigid and especially when they are enforced by software (rating system filtering) a human’s tendency is to try and satisfy it technically and not in spirit
The problem here is the same as the problem of bad imports (large magnitude of potentially bad data), so a similar solution should help (set a precedence to get an informal approval from the community before starting a challenge)
DISCLAIMER: I have never used MapRoulette before, so kindly ignore this message if the idea sounds like nonsense
I’d like to start with basic usability here. I zoom in to my city and see a bunch of green splotches with the number two in them with no way to zoom in further to discern them, and a list of possible problems to fix with no idea how many of each there are. (Please fix both of these.) Also if there are five types of errors and I know that fixing one of them requires me to actually go there, I’d like to de-select / hide that one type. (Please add a “hide” button to each.)
Next, I click on the one that’s of interest I’m immediately zoomed out to the whole of Germany or whatever region the problem covers. Owch. Why the heck do you do that? to me that makes no sense whatsoever. Please turn that non-feature off!
OK now I can zoom back in until I see individual markers. Let’s say there’s a crossing with five of those next to each other. I want to go and edit all of them at once, there’s no point in opening the editor five times on the same spot. Except there’s no way to open iD with all ways / nodes with the problem highlighted or tagged in some way, so I have to switch back and forth between views. (I admit that implementing a way to tell iD to keep a highlight on multiple ways is probably out of scope for this thread …)
After I have completed that, I go back to MR and mark whichever marker as done. Which is fine except that now I get to say “done” four more times, assuming that the “nearby” thing actually gets all five. If the markers in question are more spread out, I’ve seen it go off to a tangent halfway through.
Next, when I’m done with that chore I don’t want to go to the next-closest single node. Instead, I want to find the next cluster and work on that – the trivial reason is that opening iD takes long enough already, but also because often I want to see the global pattern to prevent mis-labelling things.
Thanks for sharing @Smurf ! All of the above are on my radar.
Map usability could be much better. I haven’t given that too much attention. I am considering switching from leaflet to MapLibre - some testing suggests that it’s faster when displaying large numbers of points (tasks) and filtering / clustering. Also with a vector tile background you’re not as limited in how far you can zoom in. I agree that you should be able to zoom in far enough that you could see all individual tasks without clustering.
A lot of your concerns fall in the domain of discoverabililty. Some of the things that frustrate you actually are possible to avoid, but the options / settings are too hard to find and it’s not intuitive enough. (For example, we have Task Bundling that lets you take multiple tasks at once and fix them all in one session. Not perfect but it works really well for some people.)
Regarding your comment about iD: I think we should actually have a way for iD to receive multiple targets to select. JOSM remote control can do this, I think it could be built without breaking the current URL parameter contract. Instead of just one id=n12345 you could select multiple features like w179388095,n8863865910. Rapid supports this already (example). @tyr_asd
In general though, the context switching between MR and editor is felt by many users as painful. We added an ‘edit mode’ that uses a built in Rapid editor to prevent having to switch to a different window or tab - it’s not perfect but it’s a start.
I’ve written down a few ideas. I apologize if these have already been mentioned on GitHub or similar; to be honest, I didn’t check for duplicates. I translated the functions in MR blindly from German to English, so they may be called something different in the original. Thank you for your great work! If I notice anything else while working with MR, I’ll submit it later.
Pain point: The data is now over a year old, and much of it has already been updated in OSM. A lot of time is wasted realizing that it is already correct in OSM.
Solution: Either automatic refreshing or the option to trigger refreshing of tasks even as a non-owner of the challenge.
Even better: This would also work for GeoJSON challenges.
Limited challenge search by location/region
User story: I only want to see challenges that are exactly within the boundary of a place or an admin relation so that I can work on tasks there specifically.
Pain point: Currently, a lot of challenges are displayed in the list on the left that are not even in the Bbox.
Solution: You specify a specific region (area of an admin relation) and only challenges that are within it are displayed.
Hiding challenges
User story: I want to be able to hide challenges if I don’t like their content / the tasks.
Pain point: You often see the same challenges that aim to improve global “minor issues”, for example, but these are too complicated or irrelevant for me.
Solution: You can hide these challenges for your own user so that they are no longer displayed in the search. (Similar to hiding quests in StreetComplete)
Filtering by “open tasks only”
User story: I would like to be able to filter the search/map to show only challenges that contain “open” tasks, so that “difficult” or “skipped” tasks are ignored.
Pain point: Often, tasks in your area are marked as “too difficult,” and then the entire challenge still appears in the list.
Solution: A checkbox to filter only ‘open’ challenges, which then hides tasks marked as “too difficult” or “skipped.”
Deactivate “Please confirm” dialog when completing a task
User story: As a user, I don’t want to see the “Please confirm” dialog when completing a task so that I can work faster.
Pain point: In 99% of cases, no comment or MR tag is entered upon completion, yet the dialog appears every time, requiring just one more click.
Solution: As a user, I want to be able to disable this dialog globally in MR.
Notification for new challenges in an area
User story: I would like to be notified when there are new tasks in an area I have specified so that I can actively work on them.
Pain point: I have to regularly navigate the map and try to manually identify whether new challenges/tasks have been added.
Solution: I would like to be able to specify an admin-related area and receive a notification as soon as new tasks are created in that area. (Optionally adjustable whether in MR or even by email)
@mcliquid first of all thanks for your continued active participation in improving MapRoulette. We’ve had quite a few constructive discussions based on your contributions to the Github issues.
We do have a mechanism that makes challenges with no activity invisible after six months. Perhaps it doesn’t work as intended right now.. Will need to look into this. Agree that it’s a major issue. Challenge authors should not just dump tasks and never look at it again. A lot of things we can do to improve here, I think it’s crucial that we get this right.
The central topic of my original question opening this thread. How can we make discovery better? I am asking a few questions about this in the survey I posted a couple of days ago to find out what the specific common pain points are with it.
Yes, mentioned by a few mappers here already. I agree that a more personalized experience would be useful (saved filters or searches, hiding challenges just for yourself, things like that, open to more ideas here)
Yes, perhaps part of the same topic as above. Being able to set your preferences for the individual ‘lens’ on the tasks you see and don’t see.
I’ve heard that asked a few times over the years. I think it could be a user setting.
If you do any rating system at all, start out with binary like/dislike - this is where the megacorps and research has been pointing towards. Use a confidence interval calculation so that +10000/-5 is higher rated than +1/-0.
Once that’s already in place, if you want to go more complicated, I’d suggest the “balanced 5-point” rating system: -2 (Something is extremely wrong and this should be removed), -1 (I don’t like it), 0 (no opinion / acceptable / average), +1 (good / i like it), +2 (Extremely good, this should be featured as an example of how to do things on this site)
(Something is extremely wrong and this should be removed), -1 
(I don’t like it), 0 
(no opinion / acceptable / average), +1 
(good / i like it), +2 
(Extremely good, this should be featured as an example of how to do things on this site)