OsmAnd Live still uses deleted street

[SchLois]

Hello,

three days ago I reported, that an existing street had been removed (https://www.openstreetmap.org/note/2328577). This change has been updated very fast on 2020-08-28 and is visible in this changeset https://www.openstreetmap.org/changeset/90140820#map=19/48.17355/11.81851 .

I updated the OsmAnd Live map several times – now version 2020-09-02 11:05 is on my mobile. But the removed street is still existent in OsmAnd.

I deleted the live update files in /storage/0000-0000/Android/data/net.osmand.plus/live and forced a new live update, but the street still seems to exist.

OsmAnd Live works in general, new versions are found and downloaded. A change on an POI nearby that I reported only with a view minutes difference to the removed street report is visible in OsmAnd.

Is there anything I can do, or is this a bug in OsmAnd Live?

I'm using OsmAnd+ Version 3.7.4 with activated "OsmAnd-Live-Navigation"on Android 7.0 Kernel 3.10.61-12236002.

[Tom Crocker]

When you force update osmand live it tells you when the last map update was (as well as the time you did the update). Usually that's within an hour or so but occasionally it gets behind. When does it say for your update (after/ before the change)?

--
You received this message because you are subscribed to the Google Groups "OsmAnd" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/osmand/8da9580a-b56b-4604-9251-8d5cbbf20050o%40googlegroups.com.

[SchLois]

The last change of the map was on 2 Sep 2020 11:05. The street was removed days before in openstreetmap.org, the poi change reported and fixed on the same day like the street removal is visible in OsmAnd Live.

[Tom Crocker]

The roads haven't been deleted, they've been retagged as 

     removed:highway=residential

However, given that, they probably still shouldn't show up. There's a similar situation near me with a historic quarry site that shows as a current quarry.

To view this discussion on the web visit https://groups.google.com/d/msgid/osmand/90d72a69-6f4a-4e62-a6db-c2480b816f64n%40googlegroups.com.

[Xavier]

The toplevel tag name "removed" is not documented in the OSM Wiki, so
that is not a proper way to "remove" a street from OpenStreetMap, which
explains why it did not disappear from OsmAnd Live.

If the street really has been removed (as in no longer exists) then the
portion of the way in OSM that corresponds to the removed street should
simply be deleted from OSM.

On Wed, Sep 02, 2020 at 06:54:17PM +0100, Tom Crocker wrote:
>The roads haven't been deleted, they've been retagged as
> removed:highway=residential
>
>However, given that, they probably still shouldn't show up. There's a
>similar situation near me with a historic quarry site that shows as a
>current quarry.
>
>
>On Wed, 2 Sep 2020, 14:11 SchLois, <alois....@gmail.com> wrote:
>
>> The last change of the map was on 2 Sep 2020 11:05. The street was removed
>> days before in openstreetmap.org, the poi change reported and fixed on
>> the same day like the street removal is visible in OsmAnd Live.
>> [image: Screenshot_20200902-123307.png]
>> tomcroc...@gmail.com schrieb am Mittwoch, 2. September 2020 um 14:13:38
>> UTC+2:
>>
>>> When you force update osmand live it tells you when the last map update
>>> was (as well as the time you did the update). Usually that's within an hour
>>> or so but occasionally it gets behind. When does it say for your update
>>> (after/ before the change)?
>>>
>>>
>>> On Wed, 2 Sep 2020, 13:05 SchLois, <alois....@gmail.com> wrote:
>>>
>>>> Hello,
>>>>
>>>> three days ago I reported, that an existing street had been removed (
>>>> https://www.openstreetmap.org/note/2328577). This change has been
>>>> updated very fast on 2020-08-28 and is visible in this changeset
>>>> https://www.openstreetmap.org/changeset/90140820#map=19/48.17355/11.81851
>>>> .
>>>>
>>>>
>>>>

>>>> I updated the OsmAnd Live map several times ??? now version 2020-09-02

>>>> 11:05 is on my mobile. But the removed street is still existent in OsmAnd.
>>>>
>>>>
>>>> I deleted the live update files in
>>>> /storage/0000-0000/Android/data/net.osmand.plus/live and forced a new
>>>> live update, but the street still seems to exist.
>>>>
>>>>

[Xavier]

Oops, my search of the OSM wiki failed me, there is indeed a "removed"
toplevel namespace, but its usage is to label particular individual
keys as removed from an object, not to indicate the object as a whole
was removed.

So this tagging would say that the "highway=residential" tag was
removed, not that the way itself was removed.

So this implies that OsmAnd's processor of OSM data is likely not
programmed to recognize the removed: namespace for indicating that a
road no longer exists. If the road is truly gone, it would still be
better to simply delete the portion of the way that corresponds than to
try to label it "removed:highway". The prior data will live on in the
OSM history in any case.

[Tom Crocker]

It's in the lifecycle prefixes page alongside demolished: and razed:. If it's only just been removed I think it makes sense to mark it with one of those prefixes so it isn't re-added based on imagery.

--
You received this message because you are subscribed to the Google Groups "OsmAnd" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/osmand/20200902182804.GR9090%40d820.dp100.com.

[Greg Troxel]

"'Xavier' via OsmAnd" <osm...@googlegroups.com> writes:

> The toplevel tag name "removed" is not documented in the OSM Wiki, so
> that is not a proper way to "remove" a street from OpenStreetMap,
> which explains why it did not disappear from OsmAnd Live.

I will not argue that it is not documented, but there is no reason for
osmand or any other renderer to decide that foo:highway is the same as
highway. Unknown tags should be ignored.

> If the street really has been removed (as in no longer exists) then
> the portion of the way in OSM that corresponds to the removed street
> should simply be deleted from OSM.

Arguably true, but on the other hand people with armchair mappers
re-adding things that used to exist from imagery and having a way with a
removed tag, even if ad hoc, is a clue to those people not to re-add
them.

Regardless, OsmAnd appears incorrect here; it should have a reasoable
response to non-standard tagging

[Greg Troxel]

"'Xavier' via OsmAnd" <osm...@googlegroups.com> writes:

> So this tagging would say that the "highway=residential" tag was
> removed, not that the way itself was removed.
>
> So this implies that OsmAnd's processor of OSM data is likely not
> programmed to recognize the removed: namespace for indicating that a
> road no longer exists.

I think it's a bigger problem. If in the new data, there isn't a way
with highway=something, then in the live-updated db on one's phone,
there shouldn't be a way with highway=something. It's not crazy to
retag a way to something else, and if it turns into something that
shouldn't be rendered or routed, it shouldn't be rendered or routed in
OsmAnd.

THe kind of sync that Live is doing is very difficult, so bugs would not
surprise me. I find that it works remarkably well, although new things
are far more common than removals.

[Xavier]

On Thu, Sep 03, 2020 at 08:31:40AM -0400, Greg Troxel wrote:
>
>"'Xavier' via OsmAnd" <osm...@googlegroups.com> writes:
>
>> The toplevel tag name "removed" is not documented in the OSM Wiki,
>> so that is not a proper way to "remove" a street from OpenStreetMap,
>> which explains why it did not disappear from OsmAnd Live.
>
>I will not argue that it is not documented, but there is no reason for
>osmand or any other renderer to decide that foo:highway is the same as
>highway. Unknown tags should be ignored.

My reply to myself indicated that my first search for the key failed, I
later found the wiki page for removed:. It is at:
https://wiki.openstreetmap.org/wiki/Key:removed:

>> If the street really has been removed (as in no longer exists) then
>> the portion of the way in OSM that corresponds to the removed street
>> should simply be deleted from OSM.
>
>Arguably true, but on the other hand people with armchair mappers
>re-adding things that used to exist from imagery and having a way with
>a removed tag, even if ad hoc, is a clue to those people not to re-add
>them.

Yes. While the Wiki page does not explicitly mention this use, the use
does make sense in regards to items, such as streets, that will remain
visible in imagery for quite some time before disappearing.

>Regardless, OsmAnd appears incorrect here; it should have a reasoable
>response to non-standard tagging

The history on the key:removed page dates back to 2013, so yes,
OsmAnd's map creation code should detect this key.

In which case the OP should file a bug report on the OsmAnd issues
tracker here: https://github.com/osmandapp/OsmAnd/issues

[SchLois]

Thank you very much for your analysis. I just submitted a bug report: https://github.com/osmandapp/OsmAnd/issues/9754