Re: [Tagging] Intermittently unprotected cycle track

2019-03-29 Thread Richard Fairhurst
Thanks everyone for the comments!

althio wrote:
> My preference would be to keep the geometry, map it as a continuous
> highway=cycleway.
> For the bits without divider, I don't like protected=no however.
> I would go with no additional tagging, and more geometry (as you said:
> crossings and junctions), and let the geometry speaks.

On balance I agree and I'll go for this solution.

Please send out a search party if I haven't returned in three days from the
maze of nested relations that is cycle routes in East London.

cheers
Richard



--
Sent from: http://gis.19327.n8.nabble.com/Tagging-f5258744.html

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] Intermittently unprotected cycle track

2019-03-27 Thread Volker Schmidt
On Wed, 27 Mar 2019, 16:14 althio,  wrote:

> Route relation membership cannot be clearly and separately applied
> with namespacing, it requires a separate objet [1].
>

I do not see any problem with that. I have been including roads with
cycleway=lane routinely in bicycle routes.

>
>
> ... or ...
> ... maybe ...
> ... I don't know if I should ...
> Apparently highway=cycleway + cycleway=lane is already in use
> 1500 uses in
> https://taginfo.openstreetmap.org/tags/cycleway=lane#combinations
> So either some people are mistagging, or the wiki needs to be updated
> to the practice
>
This looks very wrong to me. It indicates a cycling lane on a cycleway.
When I encounter this tagging I change it.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] Intermittently unprotected cycle track

2019-03-27 Thread althio
> This is correctish in terms of tagging but not in terms of geometry.
> [...] Breaking geometry to enable tagging is bad in itself, misleading on 
> renderings, and unsurprisingly confuses the heck out of routers.

Indeed. Either as cycleway=track/lane on car road (all along) or as a
separate way (all along), both not like this.

For me, adding geometry gives additional information and accuracy,
more clarity + "easily apply tagging that applies distinctly to the
cycleway (surface, route relation membership, etc.)"
Route relation membership cannot be clearly and separately applied
with namespacing, it requires a separate objet [1].
My preference would be to keep the geometry, map it as a continuous
highway=cycleway.

For the bits without divider, I don't like protected=no however.
I would go with no additional tagging, and more geometry (as you said:
crossings and junctions), and let the geometry speaks.

... or ...
... maybe ...
... I don't know if I should ...
Apparently highway=cycleway + cycleway=lane is already in use
1500 uses in https://taginfo.openstreetmap.org/tags/cycleway=lane#combinations
So either some people are mistagging, or the wiki needs to be updated
to the practice.

-- althio

[1]: if I am somehow wrong and it is indeed (remotely) possible to
apply route relation membership with namespacing, I beg, please leave
me ignorant.

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] Intermittently unprotected cycle track

2019-03-27 Thread Martin Koppenhoefer
Am Mi., 27. März 2019 um 12:39 Uhr schrieb Volker Schmidt :

> Tagging of properties of the track should be no problem:
> cycleway:left=track
> cycleway:left:width=3
> cycleway:left:surface=asphalt
> ...
>



if the cycleway is a track it can become a problem, because tracks tend to
have different properties than the road, and now you will have to split the
road also for every change of every property of a different carriageway
(the cycleway), which doesn't seem right and which makes maintenance
harder. It isn't a big problem as long as you do not map many details or
accurately map changes in properties, or if they do not vary.
It also makes it harder to understand how the position of other objects
relates to the cycleway and to the road (are they between the road and the
cycleway or further than the cycleway? etc.) and it makes it impossible to
map geometric shapes and transitions of the cycleway which are different
from the road.

Cheers,
Martin
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] Intermittently unprotected cycle track

2019-03-27 Thread Volker Schmidt
Tagging of properties of the track should be no problem:
cycleway:left=track
cycleway:left:width=3
cycleway:left:surface=asphalt
...


On Wed, 27 Mar 2019, 12:01 Tobias Zwick,  wrote:

> Hi Richard
>
> I'd tag this situation with cycleway=track/lane/shared_lane on the road
> itself. I don't see the namespacing as an issue here.
>
> The rule of thumb I (we?) use to decide whether a cycleway shall better be
> tagged as a separate way is to look if the cycleway is segregated from the
> road by more than a curb, such as scrub or a tree row. In other words, so
> that bicyclists can not simply cross onto the street at any point without a
> dedicated crossing.
>
> Tobias
>
> Am 27. März 2019 11:31:18 MEZ schrieb Richard Fairhurst
> :
> >Hi all,
> >
> >Let me introduce you to one of London's better cycleways:
> >
> >https://www.openstreetmap.org/#map=19/51.53397/-0.00715
> >https://cycle.travel/map?lat=51.5254&lon=-0.0335&zoom=17
> >
> >You might look at this and think "that doesn't look like 'better' to
> >me,
> >it's full of 45-degree bends". And based on OSM you would of course be
> >right.
> >
> >In reality it isn't full of 45-degree bends. It's a continuous straight
> >
> >route. But although it's mostly protected (i.e. concrete barrier
> >separating it from the car lanes), the protection gives out at
> >junctions
> >and crossings, so turning traffic can cross. Here's an example
> >(apologies for Google link):
> >
> >https://goo.gl/maps/rFHNHdCxMCp
> >
> >Currently, it's mapped in OSM as a highway=cycleway for the segregated
> >bits, and then it rejoins the highway=primary road (with cycleway=lane)
> >
> >where the barrier gives out.
> >
> >This is correctish in terms of tagging but not in terms of geometry.
> >The
> >current mapping implies 45-degree turns which the cyclist doesn't have
> >to take - they just continue straight on. Breaking geometry to enable
> >tagging is bad in itself, misleading on renderings, and unsurprisingly
> >confuses the heck out of routers.
> >
> >How should we represent this?
> >
> >My gut feeling is that it would be better to map it as a continuous
> >highway=cycleway but with 'protected=no' for the bits where the
> >concrete
> >divider isn't there.
> >
> >Another alternative might include deleting the cycleway completely and
> >just using cycleway=track on the car road, but this seems suboptimal as
> >
> >you can't then easily apply tagging that applies distinctly to the
> >cycleway (surface, route relation membership, etc.) without lots of
> >namespacing.
> >
> >Or we could just go with highway=cycleway and no additional tagging, on
> >
> >the basis that 'unprotected' is implied by the pedestrian-crossing tags
> >
> >and the junction geometry - i.e. obviously there's no protection there
> >because we have a junction which cars can turn across.
> >
> >Any preferences?
> >
> >cheers
> >Richard
> >
> >___
> >Tagging mailing list
> >Tagging@openstreetmap.org
> >https://lists.openstreetmap.org/listinfo/tagging
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] Intermittently unprotected cycle track

2019-03-27 Thread Tobias Zwick
Hi Richard

I'd tag this situation with cycleway=track/lane/shared_lane on the road itself. 
I don't see the namespacing as an issue here.

The rule of thumb I (we?) use to decide whether a cycleway shall better be 
tagged as a separate way is to look if the cycleway is segregated from the road 
by more than a curb, such as scrub or a tree row. In other words, so that 
bicyclists can not simply cross onto the street at any point without a 
dedicated crossing.

Tobias 

Am 27. März 2019 11:31:18 MEZ schrieb Richard Fairhurst :
>Hi all,
>
>Let me introduce you to one of London's better cycleways:
>
>https://www.openstreetmap.org/#map=19/51.53397/-0.00715
>https://cycle.travel/map?lat=51.5254&lon=-0.0335&zoom=17
>
>You might look at this and think "that doesn't look like 'better' to
>me, 
>it's full of 45-degree bends". And based on OSM you would of course be 
>right.
>
>In reality it isn't full of 45-degree bends. It's a continuous straight
>
>route. But although it's mostly protected (i.e. concrete barrier 
>separating it from the car lanes), the protection gives out at
>junctions 
>and crossings, so turning traffic can cross. Here's an example 
>(apologies for Google link):
>
>https://goo.gl/maps/rFHNHdCxMCp
>
>Currently, it's mapped in OSM as a highway=cycleway for the segregated 
>bits, and then it rejoins the highway=primary road (with cycleway=lane)
>
>where the barrier gives out.
>
>This is correctish in terms of tagging but not in terms of geometry.
>The 
>current mapping implies 45-degree turns which the cyclist doesn't have 
>to take - they just continue straight on. Breaking geometry to enable 
>tagging is bad in itself, misleading on renderings, and unsurprisingly 
>confuses the heck out of routers.
>
>How should we represent this?
>
>My gut feeling is that it would be better to map it as a continuous 
>highway=cycleway but with 'protected=no' for the bits where the
>concrete 
>divider isn't there.
>
>Another alternative might include deleting the cycleway completely and 
>just using cycleway=track on the car road, but this seems suboptimal as
>
>you can't then easily apply tagging that applies distinctly to the 
>cycleway (surface, route relation membership, etc.) without lots of 
>namespacing.
>
>Or we could just go with highway=cycleway and no additional tagging, on
>
>the basis that 'unprotected' is implied by the pedestrian-crossing tags
>
>and the junction geometry - i.e. obviously there's no protection there 
>because we have a junction which cars can turn across.
>
>Any preferences?
>
>cheers
>Richard
>
>___
>Tagging mailing list
>Tagging@openstreetmap.org
>https://lists.openstreetmap.org/listinfo/tagging

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


[Tagging] Intermittently unprotected cycle track

2019-03-27 Thread Richard Fairhurst

Hi all,

Let me introduce you to one of London's better cycleways:

https://www.openstreetmap.org/#map=19/51.53397/-0.00715
https://cycle.travel/map?lat=51.5254&lon=-0.0335&zoom=17

You might look at this and think "that doesn't look like 'better' to me, 
it's full of 45-degree bends". And based on OSM you would of course be 
right.


In reality it isn't full of 45-degree bends. It's a continuous straight 
route. But although it's mostly protected (i.e. concrete barrier 
separating it from the car lanes), the protection gives out at junctions 
and crossings, so turning traffic can cross. Here's an example 
(apologies for Google link):


https://goo.gl/maps/rFHNHdCxMCp

Currently, it's mapped in OSM as a highway=cycleway for the segregated 
bits, and then it rejoins the highway=primary road (with cycleway=lane) 
where the barrier gives out.


This is correctish in terms of tagging but not in terms of geometry. The 
current mapping implies 45-degree turns which the cyclist doesn't have 
to take - they just continue straight on. Breaking geometry to enable 
tagging is bad in itself, misleading on renderings, and unsurprisingly 
confuses the heck out of routers.


How should we represent this?

My gut feeling is that it would be better to map it as a continuous 
highway=cycleway but with 'protected=no' for the bits where the concrete 
divider isn't there.


Another alternative might include deleting the cycleway completely and 
just using cycleway=track on the car road, but this seems suboptimal as 
you can't then easily apply tagging that applies distinctly to the 
cycleway (surface, route relation membership, etc.) without lots of 
namespacing.


Or we could just go with highway=cycleway and no additional tagging, on 
the basis that 'unprotected' is implied by the pedestrian-crossing tags 
and the junction geometry - i.e. obviously there's no protection there 
because we have a junction which cars can turn across.


Any preferences?

cheers
Richard

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging