Walking & Cycling Node Network tagging: undoing the hijacking of rcn and rwn

classic Classic list List threaded Threaded
28 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

dieterdreist


sent from a phone

> On 12. Sep 2019, at 10:49, Peter Elderson <[hidden email]> wrote:
>
> If there is agreement that this actually is something worth mapping, I don't see a problem there.


this is how wikipedia works, in OpenStreetMap you do not need approval of others that something is “worth” mapping, the osm question is whether something is verifiable.


Cheers Martin
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

Janko Mihelić
I don't think this is good mapping. Firstly, this is not a route. A route is something that gets you from one place to another. This is a network of routes, and there is a tag for it, type=network[1] But this type of a relation breaks the "Relations are not Categories" rule [2]. That's why I think this network relation should be broken up into route relations with the appropriate network tag.

If this is allowed, then what stops someone from making a "Bicycle routes in Germany" relation ?



čet, 12. ruj 2019. u 11:05 Martin Koppenhoefer <[hidden email]> napisao je:


sent from a phone

> On 12. Sep 2019, at 10:49, Peter Elderson <[hidden email]> wrote:
>
> If there is agreement that this actually is something worth mapping, I don't see a problem there.


this is how wikipedia works, in OpenStreetMap you do not need approval of others that something is “worth” mapping, the osm question is whether something is verifiable.


Cheers Martin
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

dieterdreist


sent from a phone

> On 12. Sep 2019, at 11:18, Janko Mihelić <[hidden email]> wrote:
>
> I don't think this is good mapping.


agreed, I didn’t imply it was good mapping, what I said was that it can be mapped without question, because it is there, visible a verifiable.
No need for a relation at all, and I agree that it isn’t well tagged (this is not a route). A simple lcn=yes on the highways should be sufficient.

Cheers Martin
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

voschix
In reply to this post by Janko Mihelić
I see similarities of this approach with the hiking paths of the alpine clubs, but with the important difference that the routes do not have a reference.
And it's very similar to a node network, except that the nodes are not numbered.
It's a 1:1 copy of the road network signposting (and please allow the comment, has the same drawback in the sense that is not helpful to people who are not familiar with the area and don't know the names of the places and their relative positions)

I fear the only sensible thing to do is to put the lcn and REF on each way, but no relation, and map the signposts (even if there is no routing software at the moment that makes use of this information, as far as I know).
Not the best solution, but the signposting in this way does not work well either for the end user (talking from experience).

On Thu, 12 Sep 2019 at 11:21, Janko Mihelić <[hidden email]> wrote:
I don't think this is good mapping. Firstly, this is not a route. A route is something that gets you from one place to another. This is a network of routes, and there is a tag for it, type=network[1] But this type of a relation breaks the "Relations are not Categories" rule [2]. That's why I think this network relation should be broken up into route relations with the appropriate network tag.

If this is allowed, then what stops someone from making a "Bicycle routes in Germany" relation ?



čet, 12. ruj 2019. u 11:05 Martin Koppenhoefer <[hidden email]> napisao je:


sent from a phone

> On 12. Sep 2019, at 10:49, Peter Elderson <[hidden email]> wrote:
>
> If there is agreement that this actually is something worth mapping, I don't see a problem there.


this is how wikipedia works, in OpenStreetMap you do not need approval of others that something is “worth” mapping, the osm question is whether something is verifiable.


Cheers Martin
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

Peter Elderson
I think it makes sense to map preference routes as route relations, same as node2node routes within node networks. 
I am not a fan of network relations if they are just collections of elements, but if the information about how they are organised and used is also present and verifiable by survey (which is the case in the present example) it's not wrong and could be useful for maps, planners and routers.
How to use the elements in a network can be tagged by a suitable value of network:type. Currently, network_type=node_network is used, the system is increasingly adopted for all recreational transport modes. If many cities use verifiable preferential route networks, a suitable value for network:type could be added, making the route relations and network relations a system reflecting actual use rather than just collections.

A network relation without the node2node or signpost2signpost relations makes no sense to me. The signposts and arrows denote actual routes, that is the basis of the system. 

Just adding lcn=yes to the ways loses the information about the routing/network system(s) they are part of. If that's fine, no problem. If people want to map the routes and maybe the network, I think it's OK too. We have the means.

Fr gr Peter Elderson


Op do 12 sep. 2019 om 12:01 schreef Volker Schmidt <[hidden email]>:
I see similarities of this approach with the hiking paths of the alpine clubs, but with the important difference that the routes do not have a reference.
And it's very similar to a node network, except that the nodes are not numbered.
It's a 1:1 copy of the road network signposting (and please allow the comment, has the same drawback in the sense that is not helpful to people who are not familiar with the area and don't know the names of the places and their relative positions)

I fear the only sensible thing to do is to put the lcn and REF on each way, but no relation, and map the signposts (even if there is no routing software at the moment that makes use of this information, as far as I know).
Not the best solution, but the signposting in this way does not work well either for the end user (talking from experience).

On Thu, 12 Sep 2019 at 11:21, Janko Mihelić <[hidden email]> wrote:
I don't think this is good mapping. Firstly, this is not a route. A route is something that gets you from one place to another. This is a network of routes, and there is a tag for it, type=network[1] But this type of a relation breaks the "Relations are not Categories" rule [2]. That's why I think this network relation should be broken up into route relations with the appropriate network tag.

If this is allowed, then what stops someone from making a "Bicycle routes in Germany" relation ?



čet, 12. ruj 2019. u 11:05 Martin Koppenhoefer <[hidden email]> napisao je:


sent from a phone

> On 12. Sep 2019, at 10:49, Peter Elderson <[hidden email]> wrote:
>
> If there is agreement that this actually is something worth mapping, I don't see a problem there.


this is how wikipedia works, in OpenStreetMap you do not need approval of others that something is “worth” mapping, the osm question is whether something is verifiable.


Cheers Martin
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

Tagging mailing list
To summarize:
- (highway) Use lcn=yes on the highway; (my Idea) maybe with some more
Information about the network like lcn:operator=*, lcn:ref=* or similar.
- (route-relation) split up the network into smaller relations going
from guidepost to guidepost. Seems very complicated, also to query/get
the entire network.
- (network-relation) get a new value for network:type=* , maybe
guidepost_network.

I still have a slight preference toward the network-relation. It seems
very similar to the node-network from NL and imho should be tagged
comparably.

However, most import to me is to keep the information about the network
character somehow. A simple lcn=yes won't do that.

Yours
Hubert87


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

Peter Elderson
In NL node networks all node2node routes are route relations.

Then all the relations and the nodes are added to the network relation, where the network:type (i.e. the setup/system/rules), the network name, operator, website etc are tagged. Currently, the network relation for node networks is used only for maintenance en checking network integrity. 

I think the network in Bremen is a preference route system. 

Vr gr Peter Elderson


Op do 12 sep. 2019 om 22:49 schreef Hubert87 via Tagging <[hidden email]>:
To summarize:
- (highway) Use lcn=yes on the highway; (my Idea) maybe with some more
Information about the network like lcn:operator=*, lcn:ref=* or similar.
- (route-relation) split up the network into smaller relations going
from guidepost to guidepost. Seems very complicated, also to query/get
the entire network.
- (network-relation) get a new value for network:type=* , maybe
guidepost_network.

I still have a slight preference toward the network-relation. It seems
very similar to the node-network from NL and imho should be tagged
comparably.

However, most import to me is to keep the information about the network
character somehow. A simple lcn=yes won't do that.

Yours
Hubert87


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: mesh bicycle network

Tagging mailing list
Am 12.09.2019 um 23:24 schrieb Peter Elderson:
In NL node networks all node2node routes are route relations.

Then all the relations and the nodes are added to the network relation, where the network:type (i.e. the setup/system/rules), the network name, operator, website etc are tagged. Currently, the network relation for node networks is used only for maintenance en checking network integrity.
Oh, thank you for clarifying.
I think the network in Bremen is a preference route system.
Sounds about right.

Yours
Hubert87

Vr gr Peter Elderson


Op do 12 sep. 2019 om 22:49 schreef Hubert87 via Tagging <[hidden email]>:
To summarize:
- (highway) Use lcn=yes on the highway; (my Idea) maybe with some more
Information about the network like lcn:operator=*, lcn:ref=* or similar.
- (route-relation) split up the network into smaller relations going
from guidepost to guidepost. Seems very complicated, also to query/get
the entire network.
- (network-relation) get a new value for network:type=* , maybe
guidepost_network.

I still have a slight preference toward the network-relation. It seems
very similar to the node-network from NL and imho should be tagged
comparably.

However, most import to me is to keep the information about the network
character somehow. A simple lcn=yes won't do that.

Yours
Hubert87


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
12