highway=steps in bicycle profile

71 views
Skip to first unread message

Florian Lohoff

unread,
Aug 10, 2020, 7:17:54 AM8/10/20
to osm...@googlegroups.com
Hi,

last week i had the issue that OSMAnd tried to sent me through a
highway=steps with my bike. I just quickly opened an OSM Note
convinced that it must be a tagging issue:

https://www.openstreetmap.org/note/2296491

The way in question only carries a highway=steps and yes - in bicycle
profile it gets used.

A little to the east there is another steps which carries a bicycle=no
which is not used.

Does OSMAnd have some issues with steps for bikes?

Flo
--
Florian Lohoff f...@zz.de
UTF-8 Test: The 🐈 ran after a 🐁, but the 🐁 ran away
signature.asc

Greg Troxel

unread,
Aug 10, 2020, 7:28:59 AM8/10/20
to Florian Lohoff, osm...@googlegroups.com

Florian Lohoff <f...@zz.de> writes:

> Hi,
>
> last week i had the issue that OSMAnd tried to sent me through a
> highway=steps with my bike. I just quickly opened an OSM Note
> convinced that it must be a tagging issue:
>
> https://www.openstreetmap.org/note/2296491
>
> The way in question only carries a highway=steps and yes - in bicycle
> profile it gets used.
>
> A little to the east there is another steps which carries a bicycle=no
> which is not used.

That means that bicycles are prohibited.

> Does OSMAnd have some issues with steps for bikes?

It looks like they should at least be really de-preferred sort of like a
required dismount.

I would suggest reading routing.xml
signature.asc

Majka

unread,
Aug 10, 2020, 7:56:28 AM8/10/20
to OsmAnd
Not sure about your version - I am using latest from FDroid, resp. latest beta. 

You can set there the bicycle profile to avoid steps (Navigation > Settings > Avoid...)

If you cannot find it, you need to wait for the update :)

Majka

Florian Lohoff

unread,
Aug 10, 2020, 8:48:11 AM8/10/20
to Greg Troxel, osm...@googlegroups.com
Not just deprefer. Without user consent there should be NO usage of
steps AT ALL. Which would be inline with the osm wiki which
says "access=no/foot=yes" as implicit defaults.

There may be a "rail" which is okay for your city bike, not okay with
a child trailer or for me with a lot of Bikepacking gear on the bike for
a multi-week trip.

But still using a rail should be an option to select for the routing
profile not just assume that aunt tilly can carry her 30kg ebike up
steep steps on a rail.
signature.asc

Greg Troxel

unread,
Aug 10, 2020, 9:01:36 AM8/10/20
to Florian Lohoff, osm...@googlegroups.com

Florian Lohoff <f...@zz.de> writes:

> Not just deprefer. Without user consent there should be NO usage of
> steps AT ALL.

I can see that view, but there is another which is that if you want to
get from here to there on a bike, sometimes you need to carry it.
Surely it's better to do that 10m on steps than ride an extra 100 km,
for almost everyone, maybe everyone. By that logic bike routes would
never take ferries either. I meant deprefer in a very serious way,
sort of 10m of steps equivalent to maybe somewhere 2-5km of cycling.

> Which would be inline with the osm wiki which says
> "access=no/foot=yes" as implicit defaults.

In OSM, typically physical construction and access are kept somewhat
separated. I was assuming this was true, and should have checked.
It's not clear to me whether carrying a bike falls under foot access,
but given that default, I would say that steps should indeed be treated
as access=no for bicycle routing.

So I believe that osmand should be changed to treat steps as unusable
for bicycle by default.

> But still using a rail should be an option to select for the routing
> profile not just assume that aunt tilly can carry her 30kg ebike up
> steep steps on a rail.

There's another issue which is that a "30 kg ebike" is not necessarily a
"bicycle" depending on where you are. But there's alawys more
complexity in routing lurking.
signature.asc

Poutnik Fornntp

unread,
Aug 10, 2020, 9:06:57 AM8/10/20
to osm...@googlegroups.com
As an author of BRouter bike profiles, regarding access issues and fallbacks to foot access, the optimal approach is a balanced trade off:

What is the probability of (effectively ) incorrect disabling access and what is the impact ?

What is the probability of (effectively ) incorrect allowing access ( with penalty) and what is the impact ?

What is the comparison of these impacts multiplied with their probabilities ?

An optimal routing settings  depend on user cases. BRouter profile usually use high penalties for stairs, with option to disable them 

My profile use also tandem flag with extra penalties for pushing or hauling.





Dne 10. srpna 2020 14:48:10 Florian Lohoff <f...@zz.de> napsal:

-- 
You received this message because you are subscribed to the Google Groups "OsmAnd" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

Poutnik Fornntp

unread,
Aug 10, 2020, 9:11:41 AM8/10/20
to osm...@googlegroups.com, Florian Lohoff, Greg Troxel
It should be compared by the time for both activities, including mounting/unmounting, and eventually relaxing if applies.
2-5 km is typically 6-15 minutes.
Waiting above the steps for cycling mates for almost 15 minutes would be boring.

Dne 10. srpna 2020 15:01:34 Greg Troxel <g...@lexort.com> napsal:

Greg Troxel

unread,
Aug 10, 2020, 9:29:40 AM8/10/20
to Poutnik Fornntp, osm...@googlegroups.com, Florian Lohoff

Poutnik Fornntp <poutni...@gmail.com> writes:

> It should be compared by the time for both activities, including
> mounting/unmounting, and eventually relaxing if applies.
> 2-5 km is typically 6-15 minutes.
> Waiting above the steps for cycling mates for almost 15 minutes would
> be boring.

The proper metric for routing is a hard call. You are taking the
strict min-time approach, and that's certainly a valid opinion. There's
also de-preferring things that one would rather not do, which gets much
harder. Bicycle routing tends to do this more, like avoiding busy
streets.

So maybe the equivalence should be based on time, with an additional
distance to trade for not putting up with steps, if one enables them.
signature.asc

Florian Lohoff

unread,
Aug 10, 2020, 9:29:45 AM8/10/20
to osm...@googlegroups.com
Hi Greg,
This is why i say steps should be avoided completely and be made
passable by flag e.g.

"If it has a rail"
"If less than 5 steps or 2meters"
"Never"

Something like that. When i am bikepacking with my 2 kids going up
stairs is a significant hurdle as they cant carry their bikes by
lifting or even push up a rail.
Same goes for people with child trailers or aunt Tilly in the 70ies
travelling with her 30kg ebike.

Currently the cost of using steps does not seem to be significant
and people seem to tag bicycle=no to avoid osmand using it although
the osm wiki says it should be default.
signature.asc

Florian Lohoff

unread,
Aug 10, 2020, 9:31:27 AM8/10/20
to osm...@googlegroups.com, Greg Troxel
On Mon, Aug 10, 2020 at 03:11:33PM +0200, Poutnik Fornntp wrote:
> It should be compared by the time for both activities, including
> mounting/unmounting, and eventually relaxing if applies.
> 2-5 km is typically 6-15 minutes.
> Waiting above the steps for cycling mates for almost 15 minutes would be
> boring.

Its not about a time issue. Steps can be a physical barrier for some not
beeing able to carry their bike with their hands.

So as long as you not agree that "I am okay lifting my bike up stairs"
osmand should not use it.
signature.asc

Poutnik Fornntp

unread,
Aug 10, 2020, 9:49:37 AM8/10/20
to Greg Troxel, osm...@googlegroups.com, Florian Lohoff
You have not understood me correctly.

Preferences should indeed be included, like for push penalties still preferring riding at the same time. 

But penalization should not be overdone.
It applies to any penalties, like unpaved surface.

Dne 10. srpna 2020 15:29:33 Greg Troxel <g...@lexort.com> napsal:

Poutnik Fornntp

unread,
Aug 10, 2020, 9:57:06 AM8/10/20
to Greg Troxel, osm...@googlegroups.com, Florian Lohoff
Like, how many minutes of riding should be equivalent to 1 minute on stairs ? And think about scaling of this ratio too. 30 secs versus 15 minutes does not look the same as 4 mins versus 2 hours.

I do not consider cases stairs are hard show stoppers, stair can be disabled completely here.

Dne 10. srpna 2020 15:49:32 Poutnik Fornntp <poutni...@gmail.com> napsal:

Poutnik Fornntp

unread,
Aug 10, 2020, 10:02:52 AM8/10/20
to osm...@googlegroups.com, Florian Lohoff, Greg Troxel
See my other post.

Default should be IMHO the ability to overcome stairs, with reasonable penalty,
as most bikers can do that and do it frequently even out of routing scenario.

Stairs can be opted out.

But as I do not use OSMand bike routing, my opinion does not matter much. :-)

Dne 10. srpna 2020 15:31:24 Florian Lohoff <f...@zz.de> napsal:

-- 
You received this message because you are subscribed to the Google Groups "OsmAnd" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

Xavier

unread,
Aug 10, 2020, 11:29:37 AM8/10/20
to osm...@googlegroups.com
On Mon, Aug 10, 2020 at 03:29:40PM +0200, Florian Lohoff wrote:
>Hi Greg,
>
>On Mon, Aug 10, 2020 at 09:01:26AM -0400, Greg Troxel wrote:
>> Florian Lohoff <f...@zz.de> writes:
>>
>> > But still using a rail should be an option to select for the routing
>> > profile not just assume that aunt tilly can carry her 30kg ebike up
>> > steep steps on a rail.
>>
>> There's another issue which is that a "30 kg ebike" is not necessarily a
>> "bicycle" depending on where you are. But there's alawys more
>> complexity in routing lurking.
>
>This is why i say steps should be avoided completely and be made
>passable by flag e.g.
>
> "If it has a rail"
> "If less than 5 steps or 2meters"
> "Never"

OSM has tags for both number of steps (step_count tag) and handrails
(handrail:*).

But, usage is slim.

Checking taginfo: https://taginfo.openstreetmap.org/

There are:

1,086,587 highway=steps tags in OSM (and this is very likely only a
small percentage of the actual physical steps across the world).

But there are only:

95,901 step_count tags (8.83% of the highway=steps tags)

85,737 handrail tags
18,175 handrail:left tags
18,127 handrail:right tags
4,703 handrail:center tags
131 handrail:both tags

Or a total of 126,873 handrail tags (11.68% of highway=steps tags)

So while a selector flag in OsmAnd for "use if rail exists" and/or use
if "number of steps less than X" are indeed useful, there is little
data in OSM yet to allow these flags to function at the moment. So a
first order of business would be to add steps tags to OSM, and to
further add the steps_count and handrail:* tags to provide the
necessary data for the selector flags to operate upon.

Florian Lohoff

unread,
Aug 10, 2020, 5:31:23 PM8/10/20
to osm...@googlegroups.com
On Mon, Aug 10, 2020 at 11:29:31AM -0400, 'Xavier' via OsmAnd wrote:
> OSM has tags for both number of steps (step_count tag) and handrails
> (handrail:*).
>
> But, usage is slim.

Sorry - i meant ramp:bicycle=yes which is the rail to push up
your bike. With that most people will be happy except those
with a child trailer.

> So while a selector flag in OsmAnd for "use if rail exists" and/or use if
> "number of steps less than X" are indeed useful, there is little data in OSM
> yet to allow these flags to function at the moment. So a first order of
> business would be to add steps tags to OSM, and to further add the
> steps_count and handrail:* tags to provide the necessary data for the
> selector flags to operate upon.

signature.asc

Xavier

unread,
Aug 10, 2020, 5:52:56 PM8/10/20
to osm...@googlegroups.com
On Mon, Aug 10, 2020 at 11:31:17PM +0200, Florian Lohoff wrote:
>On Mon, Aug 10, 2020 at 11:29:31AM -0400, 'Xavier' via OsmAnd wrote:
>> OSM has tags for both number of steps (step_count tag) and handrails
>> (handrail:*).
>>
>> But, usage is slim.
>
>Sorry - i meant ramp:bicycle=yes which is the rail to push up
>your bike. With that most people will be happy except those
>with a child trailer.

In which case, there are only 7,358 total uses of ramp:bicycle in OSM
according to taginfo. There are only 4,667 uses of ramp:bicycle=yes.

While I do not know how many physical bicycle ramps exist world wide, I
suspect that 4,667 is but a small fraction of the physical total.

So there is likely also an opportunity here to add more data to OSM.

Reply all
Reply to author
Forward
0 new messages