osmandapp / OsmAnd

OsmAnd
https://osmand.net
Other
4.65k stars 1.02k forks source link

When final segment is not possible, display these beeline segment different dash line #17987

Open lcosmin opened 1 year ago

lcosmin commented 1 year ago

🐞 routing report

Routing engine

Routing Profile

Car

Start and end points

From: https://www.openstreetmap.org/way/1197979661 To: Bucharest ( https://www.openstreetmap.org/relation/377733 )

Actual and expected routes

Actual route

Expected route:

I would expect at least a warning. The proposed route, which involves teleportation :) , is due to the fact that the nearest road is unpaved and I accidentally had the option to avoid unpaved roads set.

And, if routing is not possible given the current "avoid" filters, then maybe mark the "teleportation" segment in some other color or maybe propose a route which doesn't respect all filters, don't know.

🌍 Your Environment

OsmAnd Version:




4.5.10

Device and Android/iOS version: Android 13

DmitryAlexei commented 1 year ago

OsmAnd~ 4.6.0#654m, released: 2023-08-22

Steps to reproduce:

Note, that starting point is located on the gravel road (see OSM)

Actual result:

Default car routing Avoid gravel enabled No warning
lcosmin commented 1 year ago

Yes, I actually noticed the whole arrow after I opened the ticket; however, this left me confused, as my first thought was that there's a bug of some sort. Now that I know about this "feature", it's different. I'd still tag it as an "UI/UX improvement" issue. However, I had this result too - notice the short arrow, like in normal navigation, but might have been from an earlier version of the app.

Could be closed in my opinion.

scaidermern commented 1 year ago

Yes, it would be very nice if OsmAnd could warn if the last or first n meters can't be calculated. Then users can double check if the calculated route is really the one they want to follow. Consider it as a safety feature.

osmonoid commented 1 year ago

Please correct me if I am wrong, but I think this issue describes a very similar situation as in #16181 Unfortunately this issue was not considered as an actual problem/bug by the developers. I still think that is in fact a serious bug and may get you into trouble when driving a car if you didn't inspect every section of your planned route beforehand (or at least the first and last part which are prone for this bug). Though not in your specific easy/bold example where finding a solution for the problem "on site" does not result in a long (kilometers) detour.

lcosmin commented 1 year ago

it sounds like the same issue, indeed.

vshcherb commented 1 year ago

For now good indication is straight line though we need to think what would be non-intrusive way to warn about "non-calculated start/end segment", actually it could happen only at start and at end. We need to be careful cause we have many offroad users which are get to used to this noncalculated start / finish segments.

osmonoid commented 1 year ago

Thank you, @vshcherb, for considering this issue.

[...] For now good indication is straight line [...]

No question/doubt about that for the two cases of

The reason why at least I, but possibly also the other users of this issue, reported it are the less obvious cases where the straight line segment is actually not identifiable/distinguishable without zooming into the problematic areas. Other users not knowing about this "unexpected, partially straight line segment route planning feature" might run into problems e.g. when driving a car.

[...] we need to think what would be non-intrusive way to warn about "non-calculated start/end segment" [...]

  • Suggestion 1 taken from issue #16181 highlight the problematic segment (e.g. using a distinctive/opposite color) current state suggestion
  • Suggestion 2
  • Once this feature is introduced and the user is confronted the first time with such a case: display a (both temporarily) dismissable info/warning which also can be turned off permanently in the navigation options.
  • Ideally, the warning contains both a shortcut/link to zoom to the problematic segment and a shortcut/link to the navigation settings sub-menu where this safety mechanism can be turned off permanently.

[...] We need to be careful cause we have many offroad users which are get to used to this noncalculated start / finish segments [...]

IMHO the suggestions are in line with these considerations.

lcosmin commented 1 year ago

This ticket resulted from my lack of knowledge about the "draw it as a straight line" feature when part of the route is inaccesible.

When the problematic segment is at the beginning of the track, rendering it in a different color would probably alert the user to the fact that there's something going on there.

However, I don't think this approach is feasible for the end of the track, as I don't think many people would visually scroll over every route they plan. Sometimes it's even difficult to do, e.g. when using Android Auto.

So, I guess some sort of "intrusive" message is necessary, like a confirmation/warning dialogue. Something that auto-dismisses itself, like a toast notification, might go unnoticed on Android Auto for instance (because you set the destination and might start driving without looking at the car's screen).

And, obviously, the display of this warning should be configurable for those who don't want to see it.

scaidermern commented 1 year ago

We need to be careful cause we have many offroad users which are get to used to this noncalculated start / finish segments.

Maybe there should be a separate offroad profile or some setting to make a profile offroad. In the end, offroad routing is a completely different thing than routing along roads. Sending normal users offroad (which is what OsmAnd currently does) certainly isn't the correct approach.

vshcherb commented 1 month ago

We need to redesign full screen and backend to warn about all warnings (ferry / private roads) that could happen during the route navigation

dmpr0 commented 2 weeks ago

ToDo

DmitryAlexei commented 1 week ago

Same in the iOS app https://github.com/osmandapp/OsmAnd/discussions/20996#discussioncomment-10902699