Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[riiga]

2022-12-01 23:37 skrev Minh Nguyen:

Vào lúc 02:08 2022-12-01, Volker Schmidt đã viết:
This proposal is incorrectly giving the impression that it is in the 
spirit of the crossing:markings tag.
This tag was meant to complement and refine the existing tagging of 
crossings in some cases, but certainly not to replace, wholesale the 
"crossing" key
The crossing:markings key describes the painting on the road surface, 
not the legal situation for the traffic participants, and it also 
leaves out the vertical signals (which BTW here in Italy have 
precedence over the horizontal signs in case of conflict)


The logical analogue to the successful crossing:markings=* proposal
would be a proposal for a crossing:signals=* key that introduces
options beyond what mappers can already express (in a Babelesque
manner) using the various crossing=* tags -- without impinging on the
entrenched interests behind crossing=zebra/uncontrolled/marked.

I'm working on such a proposal, in part to close a gap that mappers in
my region experience acutely. [1] Unfortunately, I didn't manage to
complete it before this more aggressive proposal went to an RfC. I
just hope this proposed deprecation doesn't poison the well for
crossing:signals=* once it's ready. The original 2019 proposal for
crossing:signals=* made the same mistake of touching the sensitive
crossing=* key.


Also what is the meaning of crossing=no?


crossing=no is for where the road geometry etc. would suggest a
pedestrian crossing but there isn't one. Some jurisdictions have
standard signs for this situation. It's probably more common in
regions where there are laws against jaywalking. If we're really
serious about deprecating crossing=*, then a more systematized tag
could be not:highway=crossing, now that the not:* prefix is fairly
well-established.
[1] 
https://wiki.openstreetmap.org/wiki/Proposed_features/Crossing_signalization


Given the feedback and comments for this proposal, I don't intend on 
moving to voting. I will support your proposal to only approve 
crossing:signals (with extensions).


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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Minh Nguyen]

Vào lúc 02:08 2022-12-01, Volker Schmidt đã viết:
This proposal is incorrectly giving the impression that it is in the 
spirit of the crossing:markings tag.
This tag was meant to complement and refine the existing tagging of 
crossings in some cases, but certainly not to replace, wholesale the 
"crossing" key
The crossing:markings key describes the painting on the road surface, 
not the legal situation for the traffic participants, and it also leaves 
out the vertical signals (which BTW here in Italy have precedence over 
the horizontal signs in case of conflict)


The logical analogue to the successful crossing:markings=* proposal 
would be a proposal for a crossing:signals=* key that introduces options 
beyond what mappers can already express (in a Babelesque manner) using 
the various crossing=* tags -- without impinging on the entrenched 
interests behind crossing=zebra/uncontrolled/marked.


I'm working on such a proposal, in part to close a gap that mappers in 
my region experience acutely. [1] Unfortunately, I didn't manage to 
complete it before this more aggressive proposal went to an RfC. I just 
hope this proposed deprecation doesn't poison the well for 
crossing:signals=* once it's ready. The original 2019 proposal for 
crossing:signals=* made the same mistake of touching the sensitive 
crossing=* key.



Also what is the meaning of crossing=no?


crossing=no is for where the road geometry etc. would suggest a 
pedestrian crossing but there isn't one. Some jurisdictions have 
standard signs for this situation. It's probably more common in regions 
where there are laws against jaywalking. If we're really serious about 
deprecating crossing=*, then a more systematized tag could be 
not:highway=crossing, now that the not:* prefix is fairly well-established.
[1] 
https://wiki.openstreetmap.org/wiki/Proposed_features/Crossing_signalization


--
m...@nguyen.cincinnati.oh.us



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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Volker Schmidt]

This proposal is incorrectly giving the impression that it is in the spirit
of the crossing:markings tag.
This tag was meant to complement and refine the existing tagging of
crossings in some cases, but certainly not to replace, wholesale the
"crossing" key
The crossing:markings key describes the painting on the road surface, not
the legal situation for the traffic participants, and it also leaves out
the vertical signals (which BTW here in Italy have precedence over the
horizontal signs in case of conflict)

The statement
" As such, I propose to approve crossing:signals=* and additionally
deprecate crossing=* (except crossing=no)." is not in the spirit of the
crossing:markings wiki page
is unworkable: there are some several million crossing=* tags and it als
cannot replace the existing tagging (example: "crossing:markings=pictogram"
does not replace the tagging highway=path plus bicycle=designatet plus
foot=designated plus segregated=yes on the crossing way)

Also what is the meaning of crossing=no?

Please note that I am not saying that the actual tagging practice is good
or uniform.

Volker
(mapping cyclist in NE Italy)
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

sent from a phone

> On 30 Nov 2022, at 18:53, Mateusz Konieczny via Tagging 
>  wrote:
> 
> https://www.openstreetmap.org/node/3586404853
> (and not tagged with anything directly indicating that)


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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Mateusz Konieczny via Tagging]

30 lis 2022, 14:02 od dieterdre...@gmail.com:

> Am Mi., 30. Nov. 2022 um 01:10 Uhr schrieb Mateusz Konieczny via Tagging <> 
> tagging@openstreetmap.org> >:
>
>>
>>
>> 29 lis 2022, 22:55 od >> dieterdre...@gmail.com>> :
>>
>>>
>>>
>>> sent from a phone
>>>
 On 29 Nov 2022, at 09:02, Mateusz Konieczny via Tagging < 
 tagging@openstreetmap.org > wrote:

 "no traffic signals" applies also only in some jurisdictions

>>>
>>>
>>> If there are traffic signals the crossing in OpenStreetMap gets tagged 
>>> crossing=traffic_signals, this is regardless of jurisdiction AFAIK.
>>>
>> there are >38k crossing_ref=zebra in Poland
>> https://overpass-turbo.eu/s/1onT
>>
>
>
> yes, also here, I already wrote that the "crossing_ref" tag is about crossing 
> markings and not about typology (around here, I know that the wiki may not be 
> so clear on this). The "no traffic signals" was referring to crossing=zebra, 
> not crossing_ref
>

also 19k+ crossing=zebra, see https://overpass-turbo.eu/s/1orm
example of one with traffic lights: 
https://www.openstreetmap.org/node/3586404853
(and not tagged with anything directly indicating that)
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

Am Mi., 30. Nov. 2022 um 01:10 Uhr schrieb Mateusz Konieczny via Tagging <
tagging@openstreetmap.org>:

>
> 29 lis 2022, 22:55 od dieterdre...@gmail.com:
>
>
>
> sent from a phone
>
> On 29 Nov 2022, at 09:02, Mateusz Konieczny via Tagging <
> tagging@openstreetmap.org> wrote:
>
> "no traffic signals" applies also only in some jurisdictions
>
>
>
> If there are traffic signals the crossing in OpenStreetMap gets tagged
> crossing=traffic_signals, this is regardless of jurisdiction AFAIK.
>
> there are >38k crossing_ref=zebra in Poland
> https://overpass-turbo.eu/s/1onT
>


yes, also here, I already wrote that the "crossing_ref" tag is about
crossing markings and not about typology (around here, I know that the wiki
may not be so clear on this). The "no traffic signals" was referring to
crossing=zebra, not crossing_ref
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Mateusz Konieczny via Tagging]

29 lis 2022, 22:55 od dieterdre...@gmail.com:

>
>
> sent from a phone
>
>> On 29 Nov 2022, at 09:02, Mateusz Konieczny via Tagging 
>>  wrote:
>>
>> "no traffic signals" applies also only in some jurisdictions
>>
>
>
> If there are traffic signals the crossing in OpenStreetMap gets tagged 
> crossing=traffic_signals, this is regardless of jurisdiction AFAIK.
>
there are >38k crossing_ref=zebra in Poland
https://overpass-turbo.eu/s/1onT

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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

sent from a phone

> On 29 Nov 2022, at 11:06, Minh Nguyen  wrote:
> 
> What was the problem with crossing_ref=zebra again?


it’s applied to signal controlled crossings as well when they have zebra road 
markings.


> 
> What you seem to be suggesting is that the definition of crossing=zebra 
> should favor the regulations of some parties to the Vienna Convention over 
> other parties to the convention, let alone other countries that use the term 
> "zebra" to refer to something slightly different.


I don’t see a problem in either having slightly different implications with the 
same tag, depending on the jurisdiction where it is used, or use different tags 
for similar things (like zebra here and marked in the US).

Which would be equivalent tagging for crossing=zebra without  the crossing tag? 
I don’t see it in the proposal.

Cheers Martin 



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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

sent from a phone

> On 29 Nov 2022, at 09:02, Mateusz Konieczny via Tagging 
>  wrote:
> 
> "no traffic signals" applies also only in some jurisdictions


If there are traffic signals the crossing in OpenStreetMap gets tagged 
crossing=traffic_signals, this is regardless of jurisdiction AFAIK.

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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Peter Elderson]

In Nederland, zebra or zebra-path means just the striped pattern. No signs or 
signals are implied. It's the only named type of crossing*, everybody knows it. 
crossing=zebra is perfect for that; WYSIWIM (What You See Is What You Map). 
Very straightforward. Signals and signs may be present, but in no way do they 
define the crossing, so they have to be mapped as separate objects or 
attributes.

* Traffic experts may have many more names, just none that the general public 
knows.

Fr gr Peter Elderson

> Op 29 nov. 2022 om 11:06 heeft Minh Nguyen  het 
> volgende geschreven:
> 
> Vào lúc 23:01 2022-11-28, Martin Koppenhoefer đã viết:
 On 29 Nov 2022, at 00:52, Minh Nguyen  wrote:
>>> 
>>> Even if it weren't for iD's long-gone preset, I don't think an ostensibly 
>>> global tag should be defined based on the narrow provisions of a specific 
>>> country's laws.
>> I don’t think this is about a specific country, although it is not about all 
>> countries there are many of them that apply the concept and that seem to 
>> have decided on the feature in 1949 in an international agreement.
> 
> No zebras were harmed in the drafting of either the 1949 Geneva Protocol on 
> Road Signs and Signals [1] or the 1978 Vienna Convention on Road Signs and 
> Signals [2]. Neither treaty mentions this species by name, but the national 
> laws of some parties to the Vienna Convention do define zebra crossings.
> 
> For example, the UK requires zebra crossings to have alternating stripes as 
> well as belisha beacons. [3] Other countries, such as Vietnam, use the term 
> "zebra" specifically for the striped marking pattern 
> (crossing:markings=zebra), by contrast with two parallel lines 
> (crossing:markings=lines), but make no other provisions apart from what any 
> crossing would have. [4] Meanwhile, here in the U.S., which is not a party to 
> the convention, we walk on a distinct species of "zebra crossing" that has 
> slanted stripes. What was the problem with crossing_ref=zebra again?
> 
> What you seem to be suggesting is that the definition of crossing=zebra 
> should favor the regulations of some parties to the Vienna Convention over 
> other parties to the convention, let alone other countries that use the term 
> "zebra" to refer to something slightly different. This is unsustainable. At 
> one point, it might've been reasonable to justify the use of one national 
> definition as a historical accident, based on squatter's rights. But since 
> then, for better or worse, that definition has been overwhelmed by usage that 
> we can't characterize as cleanly.
> 
> Mappers benefit when they can be confident that others will look at their 
> tagging later on and interpret it consistent with their original intention. 
> Someone using crossing=zebra today shouldn't be under any illusion that it 
> means anything more specific than a marked crossing in practice. In that 
> light, crossing=zebra deserves to be given the same deference as 
> crossing=marked.
> 
> [1] 
> 
> [2] https://unece.org/fileadmin/DAM/trans/conventn/signalse.pdf#page=7
> [3] 
> 
> [4] 
> ,
>  p. 20; search for "vạch ngựa vằn", literally "zebra stripes"
> 
> -- 
> m...@nguyen.cincinnati.oh.us
> 
> 
> 
> ___
> 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] Feature Proposal - RFC - Crossing cleanup and deprecation

[Minh Nguyen]

Vào lúc 23:01 2022-11-28, Martin Koppenhoefer đã viết:

On 29 Nov 2022, at 00:52, Minh Nguyen  wrote:

Even if it weren't for iD's long-gone preset, I don't think an ostensibly 
global tag should be defined based on the narrow provisions of a specific 
country's laws.



I don’t think this is about a specific country, although it is not about all 
countries there are many of them that apply the concept and that seem to have 
decided on the feature in 1949 in an international agreement.


No zebras were harmed in the drafting of either the 1949 Geneva Protocol 
on Road Signs and Signals [1] or the 1978 Vienna Convention on Road 
Signs and Signals [2]. Neither treaty mentions this species by name, but 
the national laws of some parties to the Vienna Convention do define 
zebra crossings.


For example, the UK requires zebra crossings to have alternating stripes 
as well as belisha beacons. [3] Other countries, such as Vietnam, use 
the term "zebra" specifically for the striped marking pattern 
(crossing:markings=zebra), by contrast with two parallel lines 
(crossing:markings=lines), but make no other provisions apart from what 
any crossing would have. [4] Meanwhile, here in the U.S., which is not a 
party to the convention, we walk on a distinct species of "zebra 
crossing" that has slanted stripes. What was the problem with 
crossing_ref=zebra again?


What you seem to be suggesting is that the definition of crossing=zebra 
should favor the regulations of some parties to the Vienna Convention 
over other parties to the convention, let alone other countries that use 
the term "zebra" to refer to something slightly different. This is 
unsustainable. At one point, it might've been reasonable to justify the 
use of one national definition as a historical accident, based on 
squatter's rights. But since then, for better or worse, that definition 
has been overwhelmed by usage that we can't characterize as cleanly.


Mappers benefit when they can be confident that others will look at 
their tagging later on and interpret it consistent with their original 
intention. Someone using crossing=zebra today shouldn't be under any 
illusion that it means anything more specific than a marked crossing in 
practice. In that light, crossing=zebra deserves to be given the same 
deference as crossing=marked.


[1] 


[2] https://unece.org/fileadmin/DAM/trans/conventn/signalse.pdf#page=7
[3] 

[4] 
, 
p. 20; search for "vạch ngựa vằn", literally "zebra stripes"


--
m...@nguyen.cincinnati.oh.us



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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Mateusz Konieczny via Tagging]

29 lis 2022, 00:18 od dieterdre...@gmail.com:

> Crossing=zebra is about a zebra crossing, it implies also vertical signs- in 
> some jurisdictions and some conditions at least - and it implies that there 
> aren’t traffic signals. 
>
"no traffic signals" applies also only in some jurisdictions
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

sent from a phone

> On 29 Nov 2022, at 00:52, Minh Nguyen  wrote:
> 
> Even if it weren't for iD's long-gone preset, I don't think an ostensibly 
> global tag should be defined based on the narrow provisions of a specific 
> country's laws.


I don’t think this is about a specific country, although it is not about all 
countries there are many of them that apply the concept and that seem to have 
decided on the feature in 1949 in an international agreement.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Minh Nguyen]

Vào lúc 15:18 2022-11-28, Martin Koppenhoefer đã viết:

crossing:markings is just about this, road markings, and while 
crossing_ref=zebra wasn’t documented for a long time, people that added it 
around here told me it was about the presence of road markings as well.

Crossing=zebra is about a zebra crossing, it implies also vertical signs- in 
some jurisdictions and some conditions at least - and it implies that there 
aren’t traffic signals.
Neither crossing:markings nor crossing_ref (as it is applied here) say anything 
about traffic signals. Here you will usually have zebra markings on signal 
controlled crossings, but they aren’t zebra crossings of course, still 
crossing:markings=zebra applies. And many of them have the crossing_ref=zebra 
tag (I ignore this tag, it does not follow any consistent logics here, 
definitely not a tag I would want to base navigation decisions on). Maybe they 
are when the signals don’t work (not sure about it, the law here requires 
vertical signs for zebra crossings, unless at road intersections).


As you may be aware, iD used crossing=zebra for its Crosswalk preset 
between 2014 and 2019, during which the vast majority of occurrences 
were added. [1] So whatever it may have originally meant has already 
been diluted to the point that you're probably better off supplementing 
the tag with something more explicit if you care about the additional 
nuances you're describing here. In general, crossing_ref=* is supposed 
to be a localized, holistic approach to classifying crossings. If 
crossing_ref=zebra is itself too diluted, then perhaps you could use a 
different value. Unfortunately, it's a Monday, so all I can think of at 
the moment is crossing_ref=zebra2. XD


Even if it weren't for iD's long-gone preset, I don't think an 
ostensibly global tag should be defined based on the narrow provisions 
of a specific country's laws. But as you're only defending 
crossing=zebra from deprecation, rather than promoting its usage over 
less provincial tags, an alternative to deprecation would be to give up 
on crossing=* having any machine-readable meaning on its own and allow 
it to be an any-tags-you-like situation, similar to ref=* on ways.


[1] https://wiki.openstreetmap.org/wiki/Special:Diff/2396877#iD
[2] 
https://wiki.openstreetmap.org/wiki/Proposed_features/Crossing_signalization


--
m...@nguyen.cincinnati.oh.us



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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

sent from a phone

> On 28 Nov 2022, at 23:53, Minh Nguyen  wrote:
> 
> If we keep crossing=zebra around based on the argument [1] that it takes 
> fewer keystrokes or clicks than adding crossing_ref=zebra or 
> crossing:markings=zebra without using a preset, then this undermines the 
> arguments against railway=tram_crossing, railway=tram_level_crossing, and 
> probably some other contentious, de facto tags that are essentially shortcuts 
> for tagging combinations without additional semantic value.


crossing:markings is just about this, road markings, and while 
crossing_ref=zebra wasn’t documented for a long time, people that added it 
around here told me it was about the presence of road markings as well.

Crossing=zebra is about a zebra crossing, it implies also vertical signs- in 
some jurisdictions and some conditions at least - and it implies that there 
aren’t traffic signals. 
Neither crossing:markings nor crossing_ref (as it is applied here) say anything 
about traffic signals. Here you will usually have zebra markings on signal 
controlled crossings, but they aren’t zebra crossings of course, still 
crossing:markings=zebra applies. And many of them have the crossing_ref=zebra 
tag (I ignore this tag, it does not follow any consistent logics here, 
definitely not a tag I would want to base navigation decisions on). Maybe they 
are when the signals don’t work (not sure about it, the law here requires 
vertical signs for zebra crossings, unless at road intersections).
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Minh Nguyen]

Vào lúc 04:55 2022-11-28, Martin Koppenhoefer đã viết:
Just because there is a now a way to map crossing markings separate from 
other properties, does not imply all the tagging we have is not needed 
any more, rather I would see "crossing:markings" as implicit, for 
example on a "crossing=zebra".


If we keep crossing=zebra around based on the argument [1] that it takes 
fewer keystrokes or clicks than adding crossing_ref=zebra or 
crossing:markings=zebra without using a preset, then this undermines the 
arguments against railway=tram_crossing, railway=tram_level_crossing, 
and probably some other contentious, de facto tags that are essentially 
shortcuts for tagging combinations without additional semantic value.


I think the more salient question is whether the impact to muscle memory 
and software is worth the elegance of this proposal in its current 
state. Some other recent proposals have encountered difficulty because 
the tradeoff wasn't good enough. For us to make an informed decision, we 
need to know to what extent this change will affect editors, renderers, 
and routers. This touches on an ongoing discussion in the community 
forum about the additional hurdles a proposal must clear if it attempts 
to deprecate something that's in use. [2]


[1] 


[2] https://community.openstreetmap.org/t/5661

--
m...@nguyen.cincinnati.oh.us



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

Re: [Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[Martin Koppenhoefer]

Am Mo., 28. Nov. 2022 um 13:13 Uhr schrieb riiga :

> With the approval of the crossing:markings=* proposal there is now a
> satisfactory way of tagging whether a crossing is marked or not
> regardless of the crossing being uncontrolled or having traffic signals.
> For signals, there is crossing:signals=* which fulfills the same role
> but has not been formally approved yet.
>
> As such, I propose to approve crossing:signals=* and additionally
> deprecate crossing=* (except crossing=no).
>


Just because there is a now a way to map crossing markings separate from
other properties, does not imply all the tagging we have is not needed any
more, rather I would see "crossing:markings" as implicit, for example on a
"crossing=zebra".
Cheers,
Martin
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

[Tagging] Feature Proposal - RFC - Crossing cleanup and deprecation

[riiga]

Hello!

With the approval of the crossing:markings=* proposal there is now a 
satisfactory way of tagging whether a crossing is marked or not 
regardless of the crossing being uncontrolled or having traffic signals. 
For signals, there is crossing:signals=* which fulfills the same role 
but has not been formally approved yet.


As such, I propose to approve crossing:signals=* and additionally 
deprecate crossing=* (except crossing=no).


Please sound off your feedback and comments here, the corresponding OSM 
Community discussion thread or on the wiki talk page.
- 
https://wiki.openstreetmap.org/wiki/Proposed_features/Highway_crossing_cleanup
- 
https://community.openstreetmap.org/t/feature-proposal-rfc-crossing-cleanup-and-deprecation/6082


Best,
riiga

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