Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Paul Johnson]

On Tue, Jun 13, 2017 at 10:15 PM, Marc Gemis  wrote:

> But since there is no proposal for a relation, I cannot use
> it.
>

Someone want to help with that?
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Marc Gemis]

I don't understand your roundabouts example. The give way before the
roundabout can be mapped on the road entering the roundabout, not ?
What's different from another road with a give way sign ? That the
roundabout is a one-way road ? Perhaps the rules for give way signs
before roundabouts are different between the US and Europe ?

The other examples add additional requirements and were not really
what I asked for.
I think those situations even do not exist in Belgium, which does not
mean we should not be able to map them, just that we do not need a
complex tagging system for our give ways.

If we follow the KISS principle, we can still map all the give ways
where we do no have those additional requirements without using
relations and keep relations for those cases where a give way only
applies to certain transportation modes or to certain directions. You
could compare to to not mapping turn restrictions if we can map it
with a oneway-tag.

I think you will not be able to convince me that relations are a good
thing for simple cases and I will not be able to convince you of the
opposite. But since there is no proposal for a relation, I cannot use
it. So unless someone writes a proposal, relations will not be used.

regards

m.

On Tue, Jun 13, 2017 at 4:33 PM, Paul Johnson  wrote:
> On Tue, Jun 13, 2017 at 9:14 AM, Marc Gemis  wrote:
>>
>> As I asked you before, show me a real world case where you have to map
>> a give way sign on the intersection of more than 1 OSM way. After
>> mapping several hundreds of them in Belgium, I have never seen a case
>> where it is needed.
>
>
> Roundabouts.  Intersections that are signalized but marked for a right turn
> yield for bicyclists.  4-way intersections with a two-way yield.

___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Marc Gemis]

As I asked you before, show me a real world case where you have to map
a give way sign on the intersection of more than 1 OSM way. After
mapping several hundreds of them in Belgium, I have never seen a case
where it is needed.

But I'm willing to map relations as well, but then someone has to make
that proposal, move it forward, make sure the tools get updated etc.

m.

On Tue, Jun 13, 2017 at 2:19 PM, Paul Johnson  wrote:
> On Tue, Jun 13, 2017 at 7:00 AM, Marc Gemis  wrote:
>>
>> On Tue, Jun 13, 2017 at 12:56 PM, Paul Johnson 
>> wrote:
>> > How does it not?  A node doesn't infer a way that it's attached to,
>> > though a
>> > node may infer what nodes it crosses through.  This is why relations
>> > became
>> > a thing, to model things that can't be inferred from the other two
>> > primitive
>> > types, particularly for enforcement and turn restrictions.
>>
>> Are you saying that when I (or the software) traverse a way in forward
>> direction  and it encounters a node that has a give way  tag on it,
>> that it cannot read the direction tag on the node as well and then
>> decided whether it has to apply the give way or not ?
>
>
> In cases more complex than a simple one-way-in, one-way-out scenario on a
> one-way road, not with any real consistent accuracy, no.

___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Paul Johnson]

On Tue, Jun 13, 2017 at 7:00 AM, Marc Gemis  wrote:

> On Tue, Jun 13, 2017 at 12:56 PM, Paul Johnson 
> wrote:
> > How does it not?  A node doesn't infer a way that it's attached to,
> though a
> > node may infer what nodes it crosses through.  This is why relations
> became
> > a thing, to model things that can't be inferred from the other two
> primitive
> > types, particularly for enforcement and turn restrictions.
>
> Are you saying that when I (or the software) traverse a way in forward
> direction  and it encounters a node that has a give way  tag on it,
> that it cannot read the direction tag on the node as well and then
> decided whether it has to apply the give way or not ?
>

In cases more complex than a simple one-way-in, one-way-out scenario on a
one-way road, not with any real consistent accuracy, no.
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Marc Gemis]

On Tue, Jun 13, 2017 at 12:56 PM, Paul Johnson  wrote:
> How does it not?  A node doesn't infer a way that it's attached to, though a
> node may infer what nodes it crosses through.  This is why relations became
> a thing, to model things that can't be inferred from the other two primitive
> types, particularly for enforcement and turn restrictions.

Are you saying that when I (or the software) traverse a way in forward
direction  and it encounters a node that has a give way  tag on it,
that it cannot read the direction tag on the node as well and then
decided whether it has to apply the give way or not ?

m.

___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Paul Johnson]

On Mon, Jun 12, 2017 at 12:14 PM, Kevin Kenny 
wrote:

> On Mon, Jun 12, 2017 at 11:35 AM, Paul Johnson 
> wrote:
>
>> So we're ignoring that nodes don't inherit the directionality of the
>> underlying way?  Really sounds like you're trying to suggest using
>> direction=forward/backwards when a relation is what's actually needed.
>>
>
> There are already tens of thousands of STOP signs tagged with direction=*.
> The semantics don't imply that a node has a direction.
>

How does it not?  A node doesn't infer a way that it's attached to, though
a node may infer what nodes it crosses through.  This is why relations
became a thing, to model things that can't be inferred from the other two
primitive types, particularly for enforcement and turn restrictions.


> Have you even posted a proposal for the relation-based schema that you're
> advocating? You sent a link when we had this discussion before, but the
> link went to a page about traffic cameras that had no mention of STOP or
> GIVE WAY signs at all.
>

 I seem to recall there was an offer to put together the write-up out there
since my mediawiki-fu is rather awful.  If someone would like to help on
that, that'd be awesome.
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Kevin Kenny]

On Mon, Jun 12, 2017 at 11:35 AM, Paul Johnson  wrote:

> So we're ignoring that nodes don't inherit the directionality of the
> underlying way?  Really sounds like you're trying to suggest using
> direction=forward/backwards when a relation is what's actually needed.
>

There are already tens of thousands of STOP signs tagged with direction=*.
The semantics don't imply that a node has a direction. They assert 'traffic
must stop at this point when approaching the node in the given direction on
any way.'

JOSM is smart enough to warn if you reverse a way that has a STOP sign on
it. There are some guidance and navigation programs that correctly
recognize directional stop signs, with the direction tag on the node
referring to the direction of the approaching way. It appears that the
infrastructure is rapidly coming to accommodate the current tagging scheme.

You clearly don't like the practice, but it is indeed widespread. There
simply aren't a lot of STOP signs tagged at all in my area, which is why my
initial Overpass queries failed to turn up good examples when I asked the
question a couple of months ago.

Have you even posted a proposal for the relation-based schema that you're
advocating? You sent a link when we had this discussion before, but the
link went to a page about traffic cameras that had no mention of STOP or
GIVE WAY signs at all. Does any extant guidance and navigation software
recognize your schema? Do JOSM, Meerkartor or iD offer any assistance with
the appropriate tagging?
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Paul Johnson]

So we're ignoring that nodes don't inherit the directionality of the
underlying way?  Really sounds like you're trying to suggest using
direction=forward/backwards when a relation is what's actually needed.

On Mon, Jun 12, 2017 at 8:38 AM, Horea Meleg 
wrote:

> Hi all,
>
> Me and my Telenav colleagues found useful open source data for Wayne
> County regarding traffic signals and stop signs. We made a blogpost about
> this: http://blog.improve-osm.org/en/2017/06/mapping-traffic-
> signals-and-stop-signs-using-maproulette/ and created two MapRoulette
> challenges to facilitate the mapping process: one for traffic signals
> http://www.maproulette.org/view/2593 and the other for stop signs:
> http://www.maproulette.org/view/2564. Our team will start mapping traffic
> signals and everyone who is keen on mapping is welcomed to help us with any
> of the challenges.
>
> Thank you,
>
> Horea
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-us
>
>
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

[Talk-us] Mapping traffic signals and stop signs using MapRoulette

[Horea Meleg]

Hi all,
Me and my Telenav colleagues found useful open source data for Wayne County 
regarding traffic signals and stop signs. We made a blogpost about this: 
http://blog.improve-osm.org/en/2017/06/mapping-traffic-signals-and-stop-signs-using-maproulette/
 and created two MapRoulette challenges to facilitate the mapping process: one 
for traffic signals http://www.maproulette.org/view/2593 and the other for stop 
signs: http://www.maproulette.org/view/2564. Our team will start mapping 
traffic signals and everyone who is keen on mapping is welcomed to help us with 
any of the challenges.
Thank you,
Horea
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us