Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Mateusz Konieczny]

On Sat, 12 Mar 2016 18:42:25 +0530
Arun Ganesh  wrote:

> >
> > this means that all turn:lanes, change:lanes, access:lanes,
> > forward:lanes, backward:lanes, etc. should count ALL lanes
> 
> 
> Turn lanes meant for a turn in a particular direction at the
> junction. I'm not sure if the concept of a turn lane extends to a
> cycle?

My city (Kraków) is an example of place with local government
hostile toward cyclists for decades (with slight changes within last
years) and even here there are some bicycle specific turn lanes.

I know about cases of:

- bicycle only turn-lane on typical road [1][2]
- case of lane on multilane-road where other vehicles must turn right
  and cyclists are allowed to ride straight on (with other lanes
  allowing going straight on)[3]

[1]
https://www.google.com/maps/@50.0457625,19.9481423,3a,90y,3.52h,44.55t/data=!3m6!1e1!3m4!1sqIauDTBmcAOYuu6Ebyxsxg!2e0!7i13312!8i6656
[2]
https://www.google.com/maps/@50.058018,19.9251663,3a,75y,75.32h,64.72t/data=!3m6!1e1!3m4!1sdP-VrXePkNFQcusIubUVGA!2e0!7i13312!8i6656
[3]
https://www.google.com/maps/@50.090113,19.8829769,3a,75y,106.91h,112.73t/data=!3m7!1e1!3m5!1sRhy1jGObUMIRQd5P3alCAQ!2e0!6s%2F%2Fgeo1.ggpht.com%2Fcbk%3Fpanoid%3DRhy1jGObUMIRQd5P3alCAQ%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D315.81717%26pitch%3D0!7i13312!8i6656

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

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

On Sat, Mar 12, 2016 at 10:09 AM, Nicolás Alvarez  wrote:

>
> El 12 mar 2016, a las 04:18, Paul Johnson  escribió:
>
> On Fri, Mar 11, 2016 at 6:02 PM, Janko Mihelic 
> wrote:
>
>> pet, 11. ožu 2016. 21:00 Martin Koppenhoefer  je
>> napisao:
>>
>>>
>>> what about the pavement/sidewalk, shall it be included? If yes, what is
>>> a lane there, e.g. when there are (partially) physical separations?
>>>
>>>
>> No,  I wouldn't include sidewalks. You could argue that sidewalks are
>> just another lane, but there is a difference. The moment you step on a
>> street, you are following rules of a game. Everyone on the street should
>> know about each other, and act accordingly.
>> Sidewalks are a different story, you can stop and chat there, there are
>> no rules.
>>
>> So even if there are no physical separations, I would never include a
>> sidewalk in the lane scheme of a road (except with the sidewalk=*
>> attribute).
>>
>
> I'm not a fan of the sidewalk tag on the centerline way; just map the
> footways.
>
>
> I see it as simplified tagging. In Buenos Aires almost every road has
> sidewalks, and I'm surely not going to draw them all as separate ways.
> Adding the sidewalk tag is much easier. I would go as far as saying I would
> assume sidewalk=both unless specified otherwise!
>

I wouldn't.  I'm going out on a long-ass limb here, but I'm guessing that
rural highways, country roads, mountain roads, alleyways, parking aisles,
driveways, and the vast majority of ways that lack a hard surface aren't
going to have sidewalks of any kind at all.  And this would account for the
overwhelming majority of roads.  Educated guess based on the American
Rockies alone.
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

On Sat, Mar 12, 2016 at 7:12 AM, Arun Ganesh 
wrote:

> this means that all turn:lanes, change:lanes, access:lanes,
>> forward:lanes, backward:lanes, etc. should count ALL lanes
>
>
> Turn lanes meant for a turn in a particular direction at the junction. I'm
> not sure if the concept of a turn lane extends to a cycle?
>

At least in most of the US, yes.  Yes it does.  Heck, bicycle specific
turning facilities are becoming more common, too, to the point of being
routine in cities with busy bicycle infrastructure (or planning well ahead
for it).  Even if not the case, it can be helpful for routers to know which
lane might allow a vehicle but be the designated lane for a specific mode
(truck and bicycle designated lanes, for example, will tend to perform
slower than general lanes; likewise, bus and HOV lanes will tend to perform
faster).  Lane guidance need not necessarily be about the next turn
exclusively, having a lane suggestion in open traffic (for example in
Osmand, before the next turn arrow turns yellow) can be handy.  For
example, say you're driving Interstate 30 lengthwise through the DFW
metroplex.  That city has freeways that are typically four roadways,
sometimes six (and more wouldn't surprise me).  Which gets you through most
efficiently?  The carpool express road?  The expressway?  Or the local road?

Or a couple real world examples from experience:  You're on US 75 HOV
South, and you need to get to the Dallas Intercontinental Hotel in Farmer's
Branch.  Which exit do you take to get back on US 75 South, to have ample
time to make it across 5 lanes of traffic to take the best crossover to the
local lanes with access to surface streets?  As I found out the hard way,
the answer is "The first HOV exit" since the second one won't come until a
few feet after the ramp you need to take miles down the road, and I ended
up on I 685 HOV almost until Dallas International Airport and had to
backtrack a dozen miles.

Another one, and one I frequently used a CB to guide truckers through
traffic on (since they tend to hold a more consistent speed and are good at
preventing people from cutting you off; just get behind one going your way,
talk 'em through getting out of Portland, and suddenly driving the
Banfield's like riding the MAX with a stereo without having to run down
Burnside)... say you've just left I 5 North to get on I 84 East in
Portland, Oregon.  You're trying to get to the truck stops in Troutdale,
preferably with enough time left over to eat and get a shower before you
legally have to sleep before you get on the road again the next morning.
There's three lanes, and you still have 18 miles to go, 9 before the two
busiest interchanges in the state.  Traffic is manageable but quickly
squeezing together and slowing down fast.  Most truckers would go with the
middle or probably the right lane out of instinct, but this will get you
stuck in exit only lanes unannounced until the last minute, forcing lane
changes in unyielding traffic.  The correct answer would be "use the
leftmost lane and don't leave it", that lane leaves downtown Portland, and
by the time you get to Troutdale, it's the exit only lane for your truck
stop:  You only have a direct shot if you stay left on the I-5 to I 84 ramp
and stick to that lane all the way.  Knowing this saves you 30-45 minutes
travel time (and time is money).


> The turn lane data should provide enough information to a car driver on
> which lane to take for a turn. Its obviously important to know the presence
> of a cycle lane in between the car turn lanes for proper guidance, and the
> main questions is if the `turn:lane` key is suitable for capturing this
> data.
>

I'm not sure it needs to be car specific.  The lane usage needs may vary by
mode, it's important that supported modes be able to take this information
into account.  From experience, the farther away you get from a midsize
sedan or station wagon in terms of size, weight, acceleration, braking and
top speed, the more likely you're going to need to know which lane you need
to be in, and possibly how early you're either able or legally allowed to
be in it.  And bicycle left turn lanes, particularly in Portland, are often
the rightmost possible lane, swinging across all other traffic when that
lane's released, and much shorter than the general lanes.  Washington DC
has at least one bicycle facility where through traffic uses the right
lane, and all turns, left or right, uses the left lane (the bicycle lanes
are the centermost lanes, instead of the outside lanes).  People don't know
this stuff if they're not a roadgeek, and even if they are, they may not
necessarily have the exact pattern memorized for the distances required to
be efficient at knowing the pattern.
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Nicolás Alvarez]

> El 12 mar 2016, a las 04:18, Paul Johnson  escribió:
> 
>> On Fri, Mar 11, 2016 at 6:02 PM, Janko Mihelic  
>> wrote:
>> pet, 11. ožu 2016. 21:00 Martin Koppenhoefer  je 
>> napisao:
>>> 
>>> what about the pavement/sidewalk, shall it be included? If yes, what is a 
>>> lane there, e.g. when there are (partially) physical separations?
>> 
>> 
>> No,  I wouldn't include sidewalks. You could argue that sidewalks are just 
>> another lane, but there is a difference. The moment you step on a street, 
>> you are following rules of a game. Everyone on the street should know about 
>> each other, and act accordingly. 
>> Sidewalks are a different story, you can stop and chat there, there are no 
>> rules. 
>> 
>> So even if there are no physical separations, I would never include a 
>> sidewalk in the lane scheme of a road (except with the sidewalk=* 
>> attribute). 
> 
> I'm not a fan of the sidewalk tag on the centerline way; just map the 
> footways.

I see it as simplified tagging. In Buenos Aires almost every road has 
sidewalks, and I'm surely not going to draw them all as separate ways. Adding 
the sidewalk tag is much easier. I would go as far as saying I would assume 
sidewalk=both unless specified otherwise!

-- 
Nicolás___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Colin Smale]

In the Netherlands just about everything that can apply to motor vehicles can 
also be found somewhere applying to bicycles. That includes turn lanes...
--colin

On 12 March 2016 14:12:25 CET, Arun Ganesh  wrote:
>>
>> this means that all turn:lanes, change:lanes, access:lanes,
>> forward:lanes, backward:lanes, etc. should count ALL lanes
>
>
>Turn lanes meant for a turn in a particular direction at the junction.
>I'm
>not sure if the concept of a turn lane extends to a cycle?
>
>The turn lane data should provide enough information to a car driver on
>which lane to take for a turn. Its obviously important to know the
>presence
>of a cycle lane in between the car turn lanes for proper guidance, and
>the
>main questions is if the `turn:lane` key is suitable for capturing this
>data.
>
>Arun
>
>
>
>
>
>___
>talk mailing list
>talk@openstreetmap.org
>https://lists.openstreetmap.org/listinfo/talk
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Arun Ganesh]

>
> this means that all turn:lanes, change:lanes, access:lanes,
> forward:lanes, backward:lanes, etc. should count ALL lanes


Turn lanes meant for a turn in a particular direction at the junction. I'm
not sure if the concept of a turn lane extends to a cycle?

The turn lane data should provide enough information to a car driver on
which lane to take for a turn. Its obviously important to know the presence
of a cycle lane in between the car turn lanes for proper guidance, and the
main questions is if the `turn:lane` key is suitable for capturing this
data.

Arun

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

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

On Fri, Mar 11, 2016 at 6:02 PM, Janko Mihelic 
wrote:

> pet, 11. ožu 2016. 21:00 Martin Koppenhoefer  je
> napisao:
>
>>
>> what about the pavement/sidewalk, shall it be included? If yes, what is a
>> lane there, e.g. when there are (partially) physical separations?
>>
>>
> No,  I wouldn't include sidewalks. You could argue that sidewalks are just
> another lane, but there is a difference. The moment you step on a street,
> you are following rules of a game. Everyone on the street should know about
> each other, and act accordingly.
> Sidewalks are a different story, you can stop and chat there, there are no
> rules.
>
> So even if there are no physical separations, I would never include a
> sidewalk in the lane scheme of a road (except with the sidewalk=*
> attribute).
>

I'm not a fan of the sidewalk tag on the centerline way; just map the
footways.  Exception:  Painted sidewalks, where it's basically a hard
shoulder officially designated a sidewalk.  Examples of this can be found
in Tulsa on the Riverparks East Trail; where the cycleway is
single-carriageway, it's three lanes, with the westernmost lane being a
pedestrian lane, the middle lane being southbound bicycles and the
easternmost lane being northbound bicycles.
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Janko Mihelic]

pet, 11. ožu 2016. 21:00 Martin Koppenhoefer  je
napisao:

>
> what about the pavement/sidewalk, shall it be included? If yes, what is a
> lane there, e.g. when there are (partially) physical separations?
>
>
No,  I wouldn't include sidewalks. You could argue that sidewalks are just
another lane, but there is a difference. The moment you step on a street,
you are following rules of a game. Everyone on the street should know about
each other, and act accordingly.
Sidewalks are a different story, you can stop and chat there, there are no
rules.

So even if there are no physical separations, I would never include a
sidewalk in the lane scheme of a road (except with the sidewalk=*
attribute).

Janko
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Martin Koppenhoefer]

sent from a phone

> Am 11.03.2016 um 11:51 schrieb Janko Mihelić :
> 
> +1 for counting all lanes


what about the pavement/sidewalk, shall it be included? If yes, what is a lane 
there, e.g. when there are (partially) physical separations?

cheers,
Martin 
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Janko Mihelić]

+1 for counting all lanes, for whatever type of traffic. It's valuable
information for cars that bikes are driving alongside. If a router has to
choose a route for a car, a road without bikes is probably better. If a
driver has to turn left over a bike lane, router should show a danger sign,
watch out for bikes.

There are other types of traffic, like trams:

http://www.mapillary.com/map/im/FWKYU-i61lGVn0gv3aewzw

This link shows two tram lanes with parking places behind them. If a car
wants to park, it should watch out for trams from both directions, and then
cross those lanes and park. I think routers should have this information.

Janko
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

This is the more logical interpretation; the key:lanes page should be
updated.  cycleway=lane is great, but where is the lane?  What about
multiple lanes? This does happen, and they're not always next to each
other, and any location there's multiple bike lanes or bicycle turn lanes
can't be completely properly mapped without this.  This is a nagging issue
for anyplace with complex bicycle infrastructure.

On Fri, Mar 11, 2016 at 4:00 AM, Marc Gemis  wrote:

> Please read http://wiki.openstreetmap.org/wiki/Lanes, which describes
> the ":lanes"- subkey. This page states:
>
> "Please be aware that this extension covers all kinds of lanes for all
> kind of vehicles and is not restricted to motorized traffic."
>
> this means that all turn:lanes, change:lanes, access:lanes,
> forward:lanes, backward:lanes, etc. should count ALL lanes. Only the
> lanes-tag on its own count the lanes for motorized traffic.
> BTW, this is often neglected by QA tools that try to compare the sum
> of forward and backward lanes with the number in the lanes tag.
>
> regards
>
> m
>
> On Fri, Mar 11, 2016 at 10:35 AM, Abhishek Saikia
>  wrote:
> > Hi everyone,
> >
> > The definition of a lane as per
> http://wiki.openstreetmap.org/wiki/Key:lanes
> > is:
> >
> >> The lanes=* key should be used to specify the total number of traffic
> >> lanes a road has which are available to motorised traffic wider than a
> motor
> >> cycle.
> >
> > For the guide, we are extending the same definition to turn:lanes. The
> wiki
> > http://wiki.openstreetmap.org/wiki/Key:turn:lanes makes no mention of
> how to
> > incorporate cycles and non motorized vehicle into this tag. These are
> still
> > open questions that should be discussed by the community.
> >
> > Philip, if you have a particular example of how to tag turn lanes for
> buses,
> > we will incorporate it into the guide.
> >
> > Do keep giving more feedback.
> >
> > Regards,
> > Abhishek
> >
> >
> >
> > On Fri, Mar 11, 2016 at 12:52 PM, Paul Johnson 
> wrote:
> >>
> >> Really, any restricted lane should be in the total lane count and have
> the
> >> appropriate lane permissions assigned.
> >>
> >> On Thu, Mar 10, 2016 at 11:46 AM, Philip Barnes 
> >> wrote:
> >>>
> >>> On Thu, 2016-03-10 at 11:24 -0600, Paul Johnson wrote:
> >>>
> >>>
> >>>
> >>> On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia
> >>>  wrote:
> >>>
> >>> Hi Marc,
> >>>
> >>> Thank you for the kind words and prompt feedback. `turn:lanes` are
> >>> exclusively for vehicles and as per discussion with the OSRM team here:
> >>> https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at
> present
> >>> ignoring adding cycle lanes during `turn:lanes` mapping.
> >>> marc.ge...@gmail.com> wrote:
> >>>
> >>>
> >>> Bicycles are vehicles in most countries, and bike lanes do pose
> problems
> >>> for correct lane guidance, particularly in areas where these lanes are
> full
> >>> width instead of being a door-zone buffer.
> >>>
> >>>
> >>> You will also need to consider bus lanes as well.
> >>>
> >>> Phil (trigpoint)
> >>>
> >>> ___
> >>> talk mailing list
> >>> talk@openstreetmap.org
> >>> https://lists.openstreetmap.org/listinfo/talk
> >>>
> >>
> >>
> >> ___
> >> talk mailing list
> >> talk@openstreetmap.org
> >> https://lists.openstreetmap.org/listinfo/talk
> >>
> >
> >
> > ___
> > talk mailing list
> > talk@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk
> >
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Marc Gemis]

Please read http://wiki.openstreetmap.org/wiki/Lanes, which describes
the ":lanes"- subkey. This page states:

"Please be aware that this extension covers all kinds of lanes for all
kind of vehicles and is not restricted to motorized traffic."

this means that all turn:lanes, change:lanes, access:lanes,
forward:lanes, backward:lanes, etc. should count ALL lanes. Only the
lanes-tag on its own count the lanes for motorized traffic.
BTW, this is often neglected by QA tools that try to compare the sum
of forward and backward lanes with the number in the lanes tag.

regards

m

On Fri, Mar 11, 2016 at 10:35 AM, Abhishek Saikia
 wrote:
> Hi everyone,
>
> The definition of a lane as per http://wiki.openstreetmap.org/wiki/Key:lanes
> is:
>
>> The lanes=* key should be used to specify the total number of traffic
>> lanes a road has which are available to motorised traffic wider than a motor
>> cycle.
>
> For the guide, we are extending the same definition to turn:lanes. The wiki
> http://wiki.openstreetmap.org/wiki/Key:turn:lanes makes no mention of how to
> incorporate cycles and non motorized vehicle into this tag. These are still
> open questions that should be discussed by the community.
>
> Philip, if you have a particular example of how to tag turn lanes for buses,
> we will incorporate it into the guide.
>
> Do keep giving more feedback.
>
> Regards,
> Abhishek
>
>
>
> On Fri, Mar 11, 2016 at 12:52 PM, Paul Johnson  wrote:
>>
>> Really, any restricted lane should be in the total lane count and have the
>> appropriate lane permissions assigned.
>>
>> On Thu, Mar 10, 2016 at 11:46 AM, Philip Barnes 
>> wrote:
>>>
>>> On Thu, 2016-03-10 at 11:24 -0600, Paul Johnson wrote:
>>>
>>>
>>>
>>> On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia
>>>  wrote:
>>>
>>> Hi Marc,
>>>
>>> Thank you for the kind words and prompt feedback. `turn:lanes` are
>>> exclusively for vehicles and as per discussion with the OSRM team here:
>>> https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at present
>>> ignoring adding cycle lanes during `turn:lanes` mapping.
>>> marc.ge...@gmail.com> wrote:
>>>
>>>
>>> Bicycles are vehicles in most countries, and bike lanes do pose problems
>>> for correct lane guidance, particularly in areas where these lanes are full
>>> width instead of being a door-zone buffer.
>>>
>>>
>>> You will also need to consider bus lanes as well.
>>>
>>> Phil (trigpoint)
>>>
>>> ___
>>> talk mailing list
>>> talk@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>>
>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>

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

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Abhishek Saikia]

Hi everyone,

The definition of a lane as per http://wiki.openstreetmap.org/wiki/Key:lanes
is:

> The *lanes*=* key should be used to specify the total number of traffic
lanes  a road has which are available
to *motorised traffic* wider than a motor cycle.

For the guide, we are extending the same definition to turn:lanes. The wiki
http://wiki.openstreetmap.org/wiki/Key:turn:lanes makes no mention of how
to incorporate cycles and non motorized vehicle into this tag. These are
still open questions that should be discussed by the community.

Philip, if you have a particular example of how to tag turn lanes for
buses, we will incorporate it into the guide.

Do keep giving more feedback.

Regards,
Abhishek



On Fri, Mar 11, 2016 at 12:52 PM, Paul Johnson  wrote:

> Really, any restricted lane should be in the total lane count and have the
> appropriate lane permissions assigned.
>
> On Thu, Mar 10, 2016 at 11:46 AM, Philip Barnes 
> wrote:
>
>> On Thu, 2016-03-10 at 11:24 -0600, Paul Johnson wrote:
>>
>>
>>
>> On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia <
>> saikia.abhi...@gmail.com> wrote:
>>
>> Hi Marc,
>>
>> Thank you for the kind words and prompt feedback. `turn:lanes` are
>> exclusively for vehicles and as per discussion with the OSRM team here:
>> https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at
>> present ignoring adding cycle lanes during `turn:lanes` mapping.
>> marc.ge...@gmail.com> wrote:
>>
>>
>> Bicycles are vehicles in most countries, and bike lanes do pose problems
>> for correct lane guidance, particularly in areas where these lanes are full
>> width instead of being a door-zone buffer.
>>
>>
>> You will also need to consider bus lanes as well.
>>
>> Phil (trigpoint)
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

Really, any restricted lane should be in the total lane count and have the
appropriate lane permissions assigned.

On Thu, Mar 10, 2016 at 11:46 AM, Philip Barnes 
wrote:

> On Thu, 2016-03-10 at 11:24 -0600, Paul Johnson wrote:
>
>
>
> On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia  > wrote:
>
> Hi Marc,
>
> Thank you for the kind words and prompt feedback. `turn:lanes` are
> exclusively for vehicles and as per discussion with the OSRM team here:
> https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at
> present ignoring adding cycle lanes during `turn:lanes` mapping.
> marc.ge...@gmail.com> wrote:
>
>
> Bicycles are vehicles in most countries, and bike lanes do pose problems
> for correct lane guidance, particularly in areas where these lanes are full
> width instead of being a door-zone buffer.
>
>
> You will also need to consider bus lanes as well.
>
> Phil (trigpoint)
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Johan C]

Hi Abhishek

1. I think it's good to have a guide. But I'm familiar with discussions
through the standard OSM ways (Wiki pages, Diaries). I find GitHub not
easily accessible. Also, the danger exists that discussions are finished on
GitHub leading to changes in the guide, however omitting changes on the
Wiki pages. I think I would prefer to have a single point of discussion in
the standard OSM way. A possibility is this page:
http://wiki.openstreetmap.org/w/index.php?title=Talk:Lane_assist=edit=1

2. Shortly after Imagic launched the lanes suffix I helped improving some
issues. However, I always missed professional input since I think inventing
a good mapping scheme for supporting a lane assist is extremely complex.
Telenav had a bit of a discussion two years ago, that was it. I think it's
good to have Mapbox involved, but where are companies like TeleNav,
GeneralMagic and NavMii in this discussion?

3. One of the difficult topics is a situation involving two junctions.
Though I like the KISS way of the :lane suffix (avoiding a relation), in
case of two junctions a relation might be unavoidable, like in this
situation:
http://map.project-osrm.org/?z=18=51.896213%2C4.540266=51.896370%2C4.539505=51.896218%2C4.541039=en
Yep, the lane assist should help the driver by showing the left lane at the
starting point. Don't know how to tag it

4. The lack of involvement of companies building navigational apps might
not help getting OSM apps to the standard of apps like TomTom. In this
situation:
http://map.project-osrm.org/?z=16=51.944695%2C4.532418=51.934661%2C4.536710=51.946632%2C4.508986=en
MagicEarth does not show any lane assistance. It probably thinks this huge
intersection is a normal highway with a normal motorway exit, which should
not be showing up. I don't know if its a programming error or a case where
we need an extra tag like: lane_assist=yes, thus helping the router to
always display the lane assist.

5. A further topic might be the definition of the lanes tag, see this diary
entry: http://www.openstreetmap.org/user/imagic/diary/28309

Cherio, Johan


2016-03-10 11:03 GMT+01:00 Abhishek Saikia :

> Hello everyone,
>
> As a part of a recent push to improve OpenStreetMap navigation data, our
> team at Mapbox started collaborating documentation to create a
> comprehensive wiki on "* Guides to improve navigation data in
> OpenStreetMap *".
> This mapping guide is still not complete and the team intends to keep on
> adding new features to the guide  and make it more comprehensive so that it
> can be a reference to anyone who wants to contribute navigation data on
> OpenStreetMap. It will be helpful and great if the community can go through
> the guide and give their valuable feedback.
>
> Regards,
>
> Abhishek
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Lester Caine]

On 10/03/16 17:46, Philip Barnes wrote:
> You will also need to consider bus lanes as well.

Complicated by selective operating times and contraflow problems :)
I got caught out by that going to SOTM at Birmingham ... OSMAND was
telling me where to turn ...

-- 
Lester Caine - G8HFL
-
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk

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

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Philip Barnes]

On Thu, 2016-03-10 at 11:24 -0600, Paul Johnson wrote:
> 
> 
> On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia 
> l.com> wrote:
> > Hi Marc,
> > 
> > Thank you for the kind words and prompt feedback. `turn:lanes` are
> > exclusively for vehicles and as per discussion with the OSRM team
> > here: https://github.com/Project-OSRM/osrm-backend/issues/2048, we
> > are at present ignoring adding cycle lanes during `turn:lanes`
> > mapping. 
> > marc.ge...@gmail.com> wrote:
> > 
> Bicycles are vehicles in most countries, and bike lanes do pose
> problems for correct lane guidance, particularly in areas where these
> lanes are full width instead of being a door-zone buffer. 
You will also need to consider bus lanes as well.
Phil (trigpoint)___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Paul Johnson]

On Thu, Mar 10, 2016 at 8:56 AM, Abhishek Saikia 
wrote:

> Hi Marc,
>
> Thank you for the kind words and prompt feedback. `turn:lanes` are
> exclusively for vehicles and as per discussion with the OSRM team here:
> https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at
> present ignoring adding cycle lanes during `turn:lanes` mapping.
> marc.ge...@gmail.com> wrote:
>

Bicycles are vehicles in most countries, and bike lanes do pose problems
for correct lane guidance, particularly in areas where these lanes are full
width instead of being a door-zone buffer.
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Mateusz Konieczny]

On Thu, 10 Mar 2016 20:26:15 +0530
Abhishek Saikia  wrote:

> `turn:lanes` are > exclusively for vehicles (...) we are at
> present ignoring adding cycle lanes during `turn:lanes` mapping.

This is inconsistent given that bicycles are vehicles.

see for example https://en.wikipedia.org/wiki/Vehicle that explicitly
lists bicycles

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

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Abhishek Saikia]

Hi Marc,

Thank you for the kind words and prompt feedback. `turn:lanes` are
exclusively for vehicles and as per discussion with the OSRM team here:
https://github.com/Project-OSRM/osrm-backend/issues/2048, we are at present
ignoring adding cycle lanes during `turn:lanes` mapping.

Please do keep sending more feedback.

Regards,
Abhishek



On Thu, Mar 10, 2016 at 7:25 PM, Marc Gemis  wrote:

> Thanks for writing this nice guide.
>
> Everything becomes more complex when there are cycle lanes involved.
> Maybe you could spend a chapter on that as well ?
> I'm thinking about turn:lanes that should take into account the cycle
> lanes, as well as permissions for the different lanes in those cases.
>
> Also dedicated bus (or psv) lanes makes the mapping harder.
>
> regards
>
> m
>
>
> On Thu, Mar 10, 2016 at 11:03 AM, Abhishek Saikia
>  wrote:
> > Hello everyone,
> >
> > As a part of a recent push to improve OpenStreetMap navigation data, our
> > team at Mapbox started collaborating documentation to create a
> comprehensive
> > wiki on " Guides to improve navigation data in OpenStreetMap". This
> mapping
> > guide is still not complete and the team intends to keep on adding new
> > features to the guide  and make it more comprehensive so that it can be a
> > reference to anyone who wants to contribute navigation data on
> > OpenStreetMap. It will be helpful and great if the community can go
> through
> > the guide and give their valuable feedback.
> >
> > Regards,
> >
> > Abhishek
> >
> > ___
> > talk mailing list
> > talk@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk
> >
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Guides to improve navigation data in OpenStreetMap

[Marc Gemis]

Thanks for writing this nice guide.

Everything becomes more complex when there are cycle lanes involved.
Maybe you could spend a chapter on that as well ?
I'm thinking about turn:lanes that should take into account the cycle
lanes, as well as permissions for the different lanes in those cases.

Also dedicated bus (or psv) lanes makes the mapping harder.

regards

m


On Thu, Mar 10, 2016 at 11:03 AM, Abhishek Saikia
 wrote:
> Hello everyone,
>
> As a part of a recent push to improve OpenStreetMap navigation data, our
> team at Mapbox started collaborating documentation to create a comprehensive
> wiki on " Guides to improve navigation data in OpenStreetMap". This mapping
> guide is still not complete and the team intends to keep on adding new
> features to the guide  and make it more comprehensive so that it can be a
> reference to anyone who wants to contribute navigation data on
> OpenStreetMap. It will be helpful and great if the community can go through
> the guide and give their valuable feedback.
>
> Regards,
>
> Abhishek
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>

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

[OSM-talk] Guides to improve navigation data in OpenStreetMap

[Abhishek Saikia]

Hello everyone,

As a part of a recent push to improve OpenStreetMap navigation data, our
team at Mapbox started collaborating documentation to create a
comprehensive wiki on "* Guides to improve navigation data in OpenStreetMap
*". This mapping
guide is still not complete and the team intends to keep on adding new
features to the guide  and make it more comprehensive so that it can be a
reference to anyone who wants to contribute navigation data on
OpenStreetMap. It will be helpful and great if the community can go through
the guide and give their valuable feedback.

Regards,

Abhishek
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk