[talk-au] Mapping each lane as a new way

Andrew Harvey andrew.harvey4 at gmail.com
Thu Oct 29 03:12:26 UTC 2020


On Tue, 27 Oct 2020 at 07:32, Nemanja Bracko (E-Search) via Talk-au <
talk-au at openstreetmap.org> wrote:

> Hi everyone,
>
>
>
> My team has recognized that there are a couple of users who are repeatedly
> making questionable modeling choices.
>
> We’ve contacted them, and while some specific issues were fixed, new ones
> kept appearing. We wanted to reach out and to hear your opinions on this
> matter.
>
>
>
> User *Supt_of_Printing
> <https://www.openstreetmap.org/user/Supt_of_Printing>*
>
> This user is not using existing ways but he adds a new way on top of the
> existing one, making a lot of duplicates. The example can be:
>
> *https://www.openstreetmap.org/way/819833424
> <https://www.openstreetmap.org/way/819833424>*
>
> There is a way that is classified as track, but just over the existing one
> he added a new unclassified road.
>

Agreed that here they shouldn't be overlapping.

>From this 2014 imagery
https://www.mapillary.com/map/im/H1Z2CCaW50MniUGV_om09Q it does look like
it's just the one road there.

Unless there is malicious intent, then best to leave a comment each time
(to give benefit of the doubt in case it's from local knowledge or survey
if it's changed on the ground recently) and if you don't hear back then fix
it up. If it turns into an edit war then you can contact the DWG email to
resolve.

The nature of OpenStreetMap is that changes do happen all the time and this
does mean you need to be constantly reviewing changes, a number of people
use OSMCha for this. Of course it's nice if we can just edit and never have
any negative edits, but the nature of OSM means this probably won't ever
happen and we have to deal with these negative changes as they happen.


> *https://www.openstreetmap.org/way/746933130
> <https://www.openstreetmap.org/way/746933130>*
>
> The other type of issue that we recognized is roundabout modeling.
>
> You can see in the provided example that way goes as a part of the
> roundabout. Also, it is noticeable that half of the roundabout has a
> [oneway] tag, but the other doesn’t.
>

Maybe it's been since fixed here but it looks correct as of now per
https://www.mapillary.com/app/?pKey=GUqIKbOVIPZKLQUUZq0Idw&focus=photo.
It's an interesting intersection that isn't a normal roundabout.


>
>
> Users *Map_baker <https://www.openstreetmap.org/user/Map_baker>* and *Supt_of_Printing
> <https://www.openstreetmap.org/user/Supt_of_Printing>*
>
> They have remodeled a couple of junctions so each lane is one way, like
> here:
>
> *https://www.openstreetmap.org/way/746563860/history
> <https://www.openstreetmap.org/way/746563860/history>*
>

Due to the edits I can't see the issue clearly with this one.


>
>
> What’s happening with this remodeling:
>
>    - Destination relations are broken;
>    - Bus routes are not connected properly;
>    - For navigation purposes additional turn restrictions would be needed
>    to avoid unpleasant turn suggestions;
>    - In some cases, roads cross each other without an intersection point;
>    - It is impossible to apply correct lane information.
>
>
>
> Another example of remodeling:
>
> *https://www.openstreetmap.org/way/789300431
> <https://www.openstreetmap.org/way/789300431>*
>
> There are no physical barriers, so it should be modeled as a single way,
> but the actual osm data is currently divided based on lane information.
>

Agreed with you here, I believe the consensus is that it should be modeled
as a single way here, so no problem fixing this by re-modelling it to use a
single way. I know it might feel pointless leaving comments on their
changesets, but if you can just to again raise the issue and explain what's
wrong and the reasons for you making your changes on top, then it makes it
easier to block them if they are still defiant as we can show that we've
provided feedback, giving opportunity for any counter points and discussion.


>
>

> Although we recognize that there are different approaches to modeling
> junctions, these users are creating a lot of issues by remodeling in this
> way.
>
> Please can you advise what to do now since I believe that all osm users
> need to spend a significant amount of time checking all their previous
> edits?
>
> Users *Map_baker <https://www.openstreetmap.org/user/Map_baker>* and *travaudat
> <https://www.openstreetmap.org/user/travaudat>*
>
> They are going across Australia and converting multiple way roundabouts
> into one. We had a discussion on Talk-AU about that, where I have tried to
> explain why it is not okay to do (since attributes may change within the
> roundabout), but I can see that practice has continued.
>

If you can point out specific changesets, and ideally once where there have
been changeset comments explaining what the issue is, then it can be
forwarded onto the DWG.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-au/attachments/20201029/79baa363/attachment.htm>


More information about the Talk-au mailing list