Re: [HOT] Mapping (also) for the routing

[john whelan]

However if you want to get to a village to deliver some sort of aid and are
on foot or donkey etc. then I assume it would be acceptable to walk through
one village to get to another? If so link using path.

Cheerio John

On 6 August 2017 at 09:10,  wrote:

> I think that the problem lies partly with the routing itself.
>
>
>
> While the mind set is about Western type infrastructure expecting a road
> to and through every village, this does not apply in other parts of the
> world.
>
>
>
> In many parts there is no fixed rule that there WILL BE a road to every
> compound/community/hamlet/village.
>
> In some places in the Amazon (and the Congo) the only access is by river.
> In the mountainous regions of Nepal the only access is by path (some pretty
> precarious). In many rural villages throughout Africa the local residents
> do not own any motorised vehicles at all and access is by path.
>
>
>
> So unless there is a road of some sort to a village or through a village
> we should not be inventing them just for the routing, the same rules that
> apply for “mapping for the render” also apply for “mapping for the router”
>
>
>
>
>
>
>
> Sent from Mail  for
> Windows 10
>
>
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapping (also) for the routing

[ralph.aytoun]

I think that the problem lies partly with the routing itself.

While the mind set is about Western type infrastructure expecting a road to and 
through every village, this does not apply in other parts of the world. 

In many parts there is no fixed rule that there WILL BE a road to every 
compound/community/hamlet/village.
In some places in the Amazon (and the Congo) the only access is by river. In 
the mountainous regions of Nepal the only access is by path (some pretty 
precarious). In many rural villages throughout Africa the local residents do 
not own any motorised vehicles at all and access is by path.

So unless there is a road of some sort to a village or through a village we 
should not be inventing them just for the routing, the same rules that apply 
for “mapping for the render” also apply for “mapping for the router”



Sent from Mail for Windows 10

___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapping (also) for the routing

[mbranco2]

Thank you Pierre for your opinion, surely I'm not an experienced
contributor for these remote (for me!) countries, and I'm glad to learn
from other mappers, but I'd also like to have a rule for the specific issue
to connect roads around hamlets/villages.
So it would be nice if you all revise/change the following:

*If two or more unclassified/track highways reach a village from
different **directions
and are not connected between, try to connect them:*

*a) if buildings/huts were not too much close one each other: let's
draw **highway(s)
(=residential) inside the village, connecting the external roads*

*b) if not, but there was at least one track near/around the village
linking **the external roads, let's map it.*

*c) if none of the above, let's draw a "logical" road outside the village
tagging **it with highway=road and fixme="confirm (routing issue)" *
*  (I think that the **definition of "highway=road" tag is appropriate for
this context, but **maybe it could be a problem for rendering and/or
routing algorithms: if so **we could tag them with "highway=track"?).*

*The last solution has to be apply only if:*
*- without that junction, routing solution should propose (at least)
several k**ilomet**ers more*
*- the ground as seen from the aerial images seems reasonably travelled by
a **4x4 vehicle*
*- the length of this junction is not so much (usually  less than 100 *
*meters)*



Mail
priva di virus. www.avast.com

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

2017-08-05 17:00 GMT+02:00 Pierre Béland :

> Mbranco
>
> This is not a simple, Let's connect road solution.
>
> This area is surrounded by main roads and it is quite difficult to
> interpret imagery in this semi-desertic area. There are probably not many
> vehicule traffic accessing these hamlets from the main roads.
>
> For the remote mappers, it is  quite difficult to interpret all the traces
> in the sand and it looks like far too many ways were added indicated as
> roads and too many connections in all directions.
>
> The responsible of this Task manager task should assure that experienced
> contributors revise this. Plus somebody from the field should look at this.
>
>
> Pierre
>
>
> --
> *De :* mbranco 
> *À :* hot@openstreetmap.org
> *Envoyé le :* jeudi 3 août 2017 20h31
> *Objet :* [HOT] Mapping (also) for the routing
>
> After the discussion [1] about routing issues inside/around hamlets and
> villages, I reviewed my mapping I've done until now in #2768 HOT project
> (Nigeria), doing the following:
>
> - if two or more unclassified/track highways reach a village from different
> directions and are not connected between, I tried to connect them:
>
> a) if buildings/huts were not too much close one each other: I drew
> highway(s) (=residential) inside the village, connecting the external roads
>
> b) if not, but there was at least one track near/around the village linking
> the external roads, I mapped it.
>
> c) if none of the above, I drew a "logical" road outside the village
> tagging
> it with highway=road and fixme="confirm (routing issue)" (I think that the
> definition of "highway=road" tag [2] is appropriate for this context, but
> maybe it could be a problem for rendering and/or routing alghoritms: if so
> we could tag them with "highway=track"?).
>
> The last solution was applied only if:
> - without that junction, routing solution should propose (at least) several
> chilometers more
> - the ground as seen from the aerial images seems reasonably travelled by a
> 4x4 vehicle
> - the lenght of this junction is not so much (almost always less than 100
> meters)
>
> I'm very interested in your opinion, my work is easily checkable with this
> overpass query [3] which extracts all the highway=road in the selected
> area:
> so, if the community doesn't agree with my solution, it's very fast to
> delete or to change tagging for all these specific highways.
>
>
> [1]
> http://gis.19327.n8.nabble.com/landuse-residential-and-
> routing-problems-tp5900081.html
> [2] http://wiki.openstreetmap.org/wiki/Tag:highway%3Droad
> [3]
> [out:xml][timeout:25];
> (way["highway"="road"](11.867,14.412,12.212,14.766););
> out meta;>;out meta qt;
>
>
>
> --
> View this message in context: http://gis.19327.n8.nabble.
> com/Mapping-also-for-the-routing-tp5900230.html
> Sent from the Humanitarian OpenStreetMap (HOT) mailing list archive at
> Nabble.com.
>
> ___
> HOT mailing list
> HOT@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
>
>
___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot

Re: [HOT] Mapping (also) for the routing

[Pierre Béland]

Mbranco
This is not a simple, Let's connect road solution.
This area is surrounded by main roads and it is quite difficult to interpret 
imagery in this semi-desertic area. There are probably not many vehicule 
traffic accessing these hamlets from the main roads. 

For the remote mappers, it is  quite difficult to interpret all the traces in 
the sand and it looks like far too many ways were added indicated as roads and 
too many connections in all directions.

The responsible of this Task manager task should assure that experienced 
contributors revise this. Plus somebody from the field should look at this.
 
Pierre 


  De : mbranco 
 À : hot@openstreetmap.org 
 Envoyé le : jeudi 3 août 2017 20h31
 Objet : [HOT] Mapping (also) for the routing
   
After the discussion [1] about routing issues inside/around hamlets and
villages, I reviewed my mapping I've done until now in #2768 HOT project
(Nigeria), doing the following:

- if two or more unclassified/track highways reach a village from different
directions and are not connected between, I tried to connect them:

a) if buildings/huts were not too much close one each other: I drew
highway(s) (=residential) inside the village, connecting the external roads

b) if not, but there was at least one track near/around the village linking
the external roads, I mapped it.

c) if none of the above, I drew a "logical" road outside the village tagging
it with highway=road and fixme="confirm (routing issue)" (I think that the
definition of "highway=road" tag [2] is appropriate for this context, but
maybe it could be a problem for rendering and/or routing alghoritms: if so
we could tag them with "highway=track"?). 

The last solution was applied only if:
- without that junction, routing solution should propose (at least) several
chilometers more
- the ground as seen from the aerial images seems reasonably travelled by a
4x4 vehicle
- the lenght of this junction is not so much (almost always less than 100
meters)

I'm very interested in your opinion, my work is easily checkable with this
overpass query [3] which extracts all the highway=road in the selected area:
so, if the community doesn't agree with my solution, it's very fast to
delete or to change tagging for all these specific highways.


[1]
http://gis.19327.n8.nabble.com/landuse-residential-and-routing-problems-tp5900081.html
[2] http://wiki.openstreetmap.org/wiki/Tag:highway%3Droad
[3] 
[out:xml][timeout:25];
(way["highway"="road"](11.867,14.412,12.212,14.766););
out meta;>;out meta qt;



--
View this message in context: 
http://gis.19327.n8.nabble.com/Mapping-also-for-the-routing-tp5900230.html
Sent from the Humanitarian OpenStreetMap (HOT) mailing list archive at 
Nabble.com.

___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot


   ___
HOT mailing list
HOT@openstreetmap.org
https://lists.openstreetmap.org/listinfo/hot