Re: [Tagging] [Talk-us] Trunk VS primary,

2020-01-06 Thread Martin Koppenhoefer
Am Mo., 6. Jan. 2020 um 06:45 Uhr schrieb Julien djakk <
djakk.geograp...@gmail.com>:

> I would vote for an importance tag, values from 1 to 6 : for some roads or
> path we could reach a cool level of details : example :
>  car:importance:commute=1, bike:importance:long-distance=3
>
> We can merge : importance=6 is for cars, bikes ... and commuting and
> long-distance (usually it is for a dead-end),
>
> Importance=5 could still be called highway=unclassified.
>


It's not the first time this is proposed ;-)
It will save you some time not to pursue the idea, rather look for former
discussion about it.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2020-01-05 Thread Julien djakk
Hello ! Please note that the highway tagging is designed for cars : there
should be also a highway-like tagging for trucks, for bikes, and for
pedestrians.

Plus : there is the commuter point of view and the long-distance point of
view :-)


I would vote for an importance tag, values from 1 to 6 : for some roads or
path we could reach a cool level of details : example :
 car:importance:commute=1, bike:importance:long-distance=3

We can merge : importance=6 is for cars, bikes ... and commuting and
long-distance (usually it is for a dead-end),

Importance=5 could still be called highway=unclassified.



Julien “djakk”


Le dim. 5 janv. 2020 à 16:46, Fernando Trebien 
a écrit :

> I know this discussion is US specific, but we've struggled with
> similar issues in Brazil as well, for very similar reasons. It seems
> we've made some progress in the southern region when we chose to judge
> importance according to a somewhat simple method (it started as: trunk
> = best routes between place=city, primary = best routes between place
> = town; then we refined the population targets for each level), with
> the cost of requiring some discussion for uncommon corner cases (such
> as when the best route between a pair of large cities actually takes
> unexpectedly undeveloped roads). Some requirements based on structure
> are still in place (primaries must be paved, motorways must be
> divided, but trunks don't have to be divided). We've also assumed that
> routing quality can only be achieved after mapping speed limits and
> surfaces and cannot depend entirely on classification. It is still an
> experimental approach, but it seems like mappers and users are much
> more satisfied now. For verifiability, after a consensus was reached,
> we documented everything in the wiki. It's a lot of work, but maybe
> something like this would work in the US as well.
>
> On Fri, Dec 27, 2019 at 3:39 PM Paul Allen  wrote:
> >
> > On Fri, 27 Dec 2019 at 17:09, yo paseopor  wrote:
> >
> >> You lost my point of view:(WHICH)  the best (or worst) conditions for a
> road you can find in a country. In some countries will be seem like a
> motorway, in other countries or zones will be a sand track. And the other
> focus: WHO can know these conditions (local communitters, people who lived
> in the country, etc.) .This is an issue OSM will have to front some day.
> And some day we will have an agreement about it.
> >
> >
> > We're actually conflating several issues:
> >
> > 1) Road construction (paved/unpaved).
> >
> > 2) Number of lanes.
> >
> > 3) Central barrier yes/no.
> >
> > 4) Entry/exit types (simple junctions/roundabouts versus motorway on/off
> ramps).
> >
> > 5) Legislation (kinds of traffic, stopping, etc).
> >
> > 6) Routeing preference:
> >
> >   a) Speed
> >   b) Distance
> >
> > In some countries, like the UK, these factors are all generally
> well-correlated.  To
> > a degree.  Good routes between important destinations tend to get good
> roads. Other
> > places, good routes between important destinations get bad roads, but
> they're still
> > the best roads around.
> >
> > I think we need to start splitting up these attributes into different
> tags and leave it
> > to editors to offer the appropriate combinations for a given country.
> Then carto can
> > handle different coutries differently.  Preferable two renderings, one
> aimed at
> > construction (motorway down to dirt track) and the other aimed at "good
> route,
> > shame about the surface."
> >
> > I now have a quote from Calvin and Hobbes going through my head: "And
> while
> > I'm dreaming, I'd like a little pony."  It's probably insoluble but if
> it is soluble
> > it will take us decades to agree on a solution.
> >
> > --
> > Paul
> >
> >
> > ___
> > Tagging mailing list
> > Tagging@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/tagging
>
>
>
> --
> Fernando Trebien
>
> ___
> 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] [Talk-us] Trunk VS primary,

2020-01-05 Thread Fernando Trebien
I know this discussion is US specific, but we've struggled with
similar issues in Brazil as well, for very similar reasons. It seems
we've made some progress in the southern region when we chose to judge
importance according to a somewhat simple method (it started as: trunk
= best routes between place=city, primary = best routes between place
= town; then we refined the population targets for each level), with
the cost of requiring some discussion for uncommon corner cases (such
as when the best route between a pair of large cities actually takes
unexpectedly undeveloped roads). Some requirements based on structure
are still in place (primaries must be paved, motorways must be
divided, but trunks don't have to be divided). We've also assumed that
routing quality can only be achieved after mapping speed limits and
surfaces and cannot depend entirely on classification. It is still an
experimental approach, but it seems like mappers and users are much
more satisfied now. For verifiability, after a consensus was reached,
we documented everything in the wiki. It's a lot of work, but maybe
something like this would work in the US as well.

On Fri, Dec 27, 2019 at 3:39 PM Paul Allen  wrote:
>
> On Fri, 27 Dec 2019 at 17:09, yo paseopor  wrote:
>
>> You lost my point of view:(WHICH)  the best (or worst) conditions for a road 
>> you can find in a country. In some countries will be seem like a motorway, 
>> in other countries or zones will be a sand track. And the other focus: WHO 
>> can know these conditions (local communitters, people who lived in the 
>> country, etc.) .This is an issue OSM will have to front some day. And some 
>> day we will have an agreement about it.
>
>
> We're actually conflating several issues:
>
> 1) Road construction (paved/unpaved).
>
> 2) Number of lanes.
>
> 3) Central barrier yes/no.
>
> 4) Entry/exit types (simple junctions/roundabouts versus motorway on/off 
> ramps).
>
> 5) Legislation (kinds of traffic, stopping, etc).
>
> 6) Routeing preference:
>
>   a) Speed
>   b) Distance
>
> In some countries, like the UK, these factors are all generally 
> well-correlated.  To
> a degree.  Good routes between important destinations tend to get good roads. 
> Other
> places, good routes between important destinations get bad roads, but they're 
> still
> the best roads around.
>
> I think we need to start splitting up these attributes into different tags 
> and leave it
> to editors to offer the appropriate combinations for a given country.  Then 
> carto can
> handle different coutries differently.  Preferable two renderings, one aimed 
> at
> construction (motorway down to dirt track) and the other aimed at "good route,
> shame about the surface."
>
> I now have a quote from Calvin and Hobbes going through my head: "And while
> I'm dreaming, I'd like a little pony."  It's probably insoluble but if it is 
> soluble
> it will take us decades to agree on a solution.
>
> --
> Paul
>
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging



-- 
Fernando Trebien

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-27 Thread Paul Allen
On Fri, 27 Dec 2019 at 17:09, yo paseopor  wrote:

You lost my point of view:(WHICH)  the best (or worst) conditions for a
> road you can find in a country. In some countries will be seem like a
> motorway, in other countries or zones will be a sand track. And the other
> focus: WHO can know these conditions (local communitters, people who lived
> in the country, etc.) .This is an issue OSM will have to front some day.
> And some day we will have an agreement about it.
>

We're actually conflating several issues:

1) Road construction (paved/unpaved).

2) Number of lanes.

3) Central barrier yes/no.

4) Entry/exit types (simple junctions/roundabouts versus motorway on/off
ramps).

5) Legislation (kinds of traffic, stopping, etc).

6) Routeing preference:

  a) Speed
  b) Distance

In some countries, like the UK, these factors are all generally
well-correlated.  To
a degree.  Good routes between important destinations tend to get good
roads. Other
places, good routes between important destinations get bad roads, but
they're still
the best roads around.

I think we need to start splitting up these attributes into different tags
and leave it
to editors to offer the appropriate combinations for a given country.  Then
carto can
handle different coutries differently.  Preferable two renderings, one
aimed at
construction (motorway down to dirt track) and the other aimed at "good
route,
shame about the surface."

I now have a quote from Calvin and Hobbes going through my head: "And while
I'm dreaming, I'd like a little pony."  It's probably insoluble but if it
is soluble
it will take us decades to agree on a solution.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-27 Thread yo paseopor
Hi!
You lost my point of view:(WHICH)  the best (or worst) conditions for a
road you can find in a country. In some countries will be seem like a
motorway, in other countries or zones will be a sand track. And the other
focus: WHO can know these conditions (local communitters, people who lived
in the country, etc.) .This is an issue OSM will have to front some day.
And some day we will have an agreement about it.

Health and maps (Salut i mapes)
yopaseopor

On Thu, Dec 26, 2019 at 7:51 PM Erkin Alp Güney 
wrote:

> No, that is highway=road. highway=unclassified is one grade above that.
>
> 26.12.2019 21:39 tarihinde yo paseopor yazdı:
>
> >
> > In a country like Zambia or Congo unclassified would be the worst
> > condition road you can find in that country (but not track), so only
> > local community people (or people who live in that country) would know
> > what are the worst conditions you can find in their country.
>
> ___
> 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] [Talk-us] Trunk VS primary,

2019-12-27 Thread Mateusz Konieczny
highway=road is solely for unsurveyed roads of unknown class
(useful for mapping from low quality aerial images).

See https://wiki.openstreetmap.org/wiki/Tag:highway%3Droad

26 Dec 2019, 19:50 by erkinalp9...@gmail.com:

> No, that is highway=road. highway=unclassified is one grade above that.
>
> 26.12.2019 21:39 tarihinde yo paseopor yazdı:
>
>>
>> In a country like Zambia or Congo unclassified would be the worst
>> condition road you can find in that country (but not track), so only
>> local community people (or people who live in that country) would know
>> what are the worst conditions you can find in their country.
>>
>
> ___
> 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] [Talk-us] Trunk VS primary,

2019-12-27 Thread Martin Koppenhoefer


sent from a phone

> On 27. Dec 2019, at 11:18, Mateusz Konieczny  wrote:
> 
> highway=road tends to be most typically used to indicate that there is a 
> traversable path of unknown quality
> +1


+1, of unknown quality, classification and access, it is a kind of fixme that 
should be modified to a specific high value if you know the situation.


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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-27 Thread Mateusz Konieczny



26 Dec 2019, 20:02 by ba...@ursamundi.org:

> On Thu, Dec 26, 2019 at 12:50 PM Erkin Alp Güney <> erkinalp9...@gmail.com> > 
> wrote:
>
>> No, that is highway=road. highway=unclassified is one grade above that.
>>
>
> highway=road tends to be most typically used to indicate that there is a 
> traversable path of unknown quality
>
+1
>
> , or a temporary road in a construction zone. 
>
I think that this is rather highway=service
(Maybe with service=driveway)
>
> These tend to be aggressively reclassified or removed as surveys happen or 
> temporary roads get removed after they're no longer needed.
>
+1___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-26 Thread Paul Johnson
On Thu, Dec 26, 2019 at 12:50 PM Erkin Alp Güney 
wrote:

> No, that is highway=road. highway=unclassified is one grade above that.
>

highway=road tends to be most typically used to indicate that there is a
traversable path of unknown quality, or a temporary road in a construction
zone.   These tend to be aggressively reclassified or removed as surveys
happen or temporary roads get removed after they're no longer needed.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-26 Thread Kevin Kenny
On Thu, Dec 26, 2019 at 1:51 PM Erkin Alp Güney  wrote:
>
> No, that is highway=road. highway=unclassified is one grade above that.

Locally to me, `highway=road` is pretty much synonymous with 'needs a
field survey'; we consider it as an interim tag for something that's
not otherwise identified. The only thing we have between
`unclassified` and `track` is `service`.

-- 
73 de ke9tv/2, Kevin

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-26 Thread Erkin Alp Güney
No, that is highway=road. highway=unclassified is one grade above that.

26.12.2019 21:39 tarihinde yo paseopor yazdı:

>
> In a country like Zambia or Congo unclassified would be the worst
> condition road you can find in that country (but not track), so only
> local community people (or people who live in that country) would know
> what are the worst conditions you can find in their country.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-26 Thread yo paseopor
First problem for classifications is the reason of the classification. What
is this reason: administrative laws (with their political facts to keep in
mind) or physical conditions (the best for the performance of the vehicles
you would have in this road)?

Second problem is the reality of the country: If you are in the country
with the biggest budget in motorways you will be sure some secondary will
have two lane roads and bridges without crossings. If you are in a country
major part of your population will not have assured every day's meal main
road would be an asphalted road or sand track.

But the number of grade classifications are the same wherever you go: four
to six grades
(1-trunk,2-primary,3-secondary,4-tertiary,5-unclassified,6-track) .
My option would be a physical classification based on the state of the
country. Only motorway would be apart of this classification because a
motorway is the same here than in other countries.

Here in Europe you can find the administrative grades at the reference of
the road (N- ,Regional- , Province- in Spain; N-, D- in France; B , L in
Austria; ...).

In a country like Spain trunk would be a road with 100 km/h medium speed
and no crossings at the same level with restriction for cycles .
In a country like Zambia or Congo trunk would be the best conditions road
you can find in that country , so only local community people (or people
who live in that country) would know what are the best conditions you can
find in their country.
In a country like Spain unclassified would be the municipality asphalted
tracks, with no more than 3 meters width (for one car).
In a country like Zambia or Congo unclassified would be the worst condition
road you can find in that country (but not track), so only local community
people (or people who live in that country) would know what are the worst
conditions you can find in their country.

I know Spain because I am Spaniard and I travel through Europe , so I think
in the European Union things would be similar. Local communitters can
explain in this list what are the best and the worst conditions for the
roads in their countries and their opinion about the other grades. I don't
have any knowledge about Zambia or Congo so I think other people would have
more knowledge than I can have about their countries.

What are your opinions? What do you think?
Salut i mapes (Health and Maps)
yopaseopor

On Mon, Dec 23, 2019 at 11:41 AM Fernando Trebien <
fernando.treb...@gmail.com> wrote:

> On Fri, 20 Dec 2019, 19:05 Graeme Fitzpatrick, 
> wrote:
>
>>
>> On Fri, 20 Dec 2019 at 19:18, Martin Koppenhoefer 
>> wrote:
>>
>>> lanes=2
>>> surface=unpaved
>>>
>>
>> But would they still count as either =trunk or =primary?
>>
>> While they're of high local importance, they're definitely not
>> high-performance & they don't link major population centres either?
>>
>
> That's exactly the problem we had in Brazil while trying to prescribe
> classification from physical characteristics from roads.
>
> It may be worth checking what are popular commercial (paper or digital)
> maps doing regarding the classification of highways and streets.
> ___
> 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] [Talk-us] Trunk VS primary,

2019-12-23 Thread Fernando Trebien
On Fri, 20 Dec 2019, 19:05 Graeme Fitzpatrick, 
wrote:

>
> On Fri, 20 Dec 2019 at 19:18, Martin Koppenhoefer 
> wrote:
>
>> lanes=2
>> surface=unpaved
>>
>
> But would they still count as either =trunk or =primary?
>
> While they're of high local importance, they're definitely not
> high-performance & they don't link major population centres either?
>

That's exactly the problem we had in Brazil while trying to prescribe
classification from physical characteristics from roads.

It may be worth checking what are popular commercial (paper or digital)
maps doing regarding the classification of highways and streets.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-22 Thread Martin Koppenhoefer


sent from a phone

> On 22. Dec 2019, at 01:24, Joseph Eisenberg  
> wrote:
> 
> Do highway=trunk in German always have a physical barrier such a kerb
> to separate the two directions, even if they are not a dual
> carriageway?


road


> 
> The English highway=trunk page says this about Germany "The
> carriageways are separated physically or by road markings".


the wiki is correct, road markings are sufficient 


> 
> An automated translation of the German page suggests that these
> "Autobahnähnliche Straße" can be translated "expressways".


from how expressways are defined in the OpenStreetMap wiki I would say they are 
pretty different on closer inspection 


> 
> But it's not clear how they are distinguished from highway=motorway
> features in Germany.


legally (motorways are signposted with start and end signs), and also 
consequently by access (trunks might be accessible by slower vehicles)


> 
> I think this shows the disadvantage of determing the top-level highway
> feature tag (primary, trunk) based on certain physical and legal
> characteristics rather than on class in the road network:


regarding the “class”, this is mostly opaque to Germans, it is quite a 
technical specialist matter, and people are generally only aware of the owner 
and maintenance entity (Bundesstraße, Landesstraße, Kreisstraße, 
Gemeindestraße) but not about the highway class, you can get an idea here 
https://de.m.wikipedia.org/wiki/Richtlinien_für_integrierte_Netzgestaltung

unfortunately the full text is not openly available.

Cheers Martin 

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Paul Johnson
On Sat, Dec 21, 2019 at 3:48 PM Martin Koppenhoefer 
wrote:

>
>
> sent from a phone
>
> > On 21. Dec 2019, at 01:10, Joseph Eisenberg 
> wrote:
> >
> > Unfortunately, the road classification system in parts of Continental
> > Europe was different, so mappers in some major countries, including
> > Germany and France, chose to use highway=trunk as synonym for
> > "motorroad" (somewhat similar to a U.S.A. "expressway"), with other
> > major roads tagged as highway=primary.
>
>
> actually not, the motorroad tag was introduced by the Germans (AFAIK) to
> express a typical access situation on many trunks but also some primaries
> (motorway like access), so that trunk (motorway like physical construction)
> and access could be tagged orthogonally. There are also some trunks that
> permit slower traffic in Germany.
>

I would also consider a "super two" or similar undivided design to be a
trunk.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Joseph Eisenberg
Do highway=trunk in German always have a physical barrier such a kerb
to separate the two directions, even if they are not a dual
carriageway?

The English highway=trunk page says this about Germany "The
carriageways are separated physically or by road markings".

An automated translation of the German page suggests that these
"Autobahnähnliche Straße" can be translated "expressways".

But it's not clear how they are distinguished from highway=motorway
features in Germany.

I think this shows the disadvantage of determing the top-level highway
feature tag (primary, trunk) based on certain physical and legal
characteristics rather than on class in the road network: a number of
different features are combined in one tag which might better be
seprate tags like "expressway=yes" + "motorroad=no" +
"dual_carriageway=yes" + "acces=" etc. to express the important
characteristics of the road.

Joseph Eisenberg

On 12/22/19, Martin Koppenhoefer  wrote:
>
>
> sent from a phone
>
>>> On 21. Dec 2019, at 22:54, Joseph Eisenberg 
>>> wrote:
>> Thank you for the correction. So highway=trunk in German is similar to
>> expressway=yes in the USA?
>
>
> I am not familiar with US tagging, but the expressway page says they must be
> dual carriageways and can have at grade intersections:
> https://wiki.openstreetmap.org/wiki/Key:expressway
>
> the German trunks must not have at grade intersections (like motorways) and
> can be single carriageway (but do not allow overtaking using the opposite
> direction).
>
>
> Cheers Martin

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Martin Koppenhoefer


sent from a phone

>> On 21. Dec 2019, at 22:54, Joseph Eisenberg  
>> wrote:
> Thank you for the correction. So highway=trunk in German is similar to 
> expressway=yes in the USA?


I am not familiar with US tagging, but the expressway page says they must be 
dual carriageways and can have at grade intersections:
https://wiki.openstreetmap.org/wiki/Key:expressway

the German trunks must not have at grade intersections (like motorways) and can 
be single carriageway (but do not allow overtaking using the opposite 
direction).


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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Joseph Eisenberg
Thank you for the correction. So highway=trunk in German is similar to
expressway=yes in the USA?

Joseph

On Sun, Dec 22, 2019 at 6:49 AM Martin Koppenhoefer 
wrote:

>
>
> sent from a phone
>
> > On 21. Dec 2019, at 01:10, Joseph Eisenberg 
> wrote:
> >
> > Unfortunately, the road classification system in parts of Continental
> > Europe was different, so mappers in some major countries, including
> > Germany and France, chose to use highway=trunk as synonym for
> > "motorroad" (somewhat similar to a U.S.A. "expressway"), with other
> > major roads tagged as highway=primary.
>
>
> actually not, the motorroad tag was introduced by the Germans (AFAIK) to
> express a typical access situation on many trunks but also some primaries
> (motorway like access), so that trunk (motorway like physical construction)
> and access could be tagged orthogonally. There are also some trunks that
> permit slower traffic in Germany.
>
> Cheers Martin
>
>
> ___
> 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] [Talk-us] Trunk VS primary,

2019-12-21 Thread Martin Koppenhoefer


sent from a phone

> On 21. Dec 2019, at 01:10, Joseph Eisenberg  
> wrote:
> 
> Unfortunately, the road classification system in parts of Continental
> Europe was different, so mappers in some major countries, including
> Germany and France, chose to use highway=trunk as synonym for
> "motorroad" (somewhat similar to a U.S.A. "expressway"), with other
> major roads tagged as highway=primary.


actually not, the motorroad tag was introduced by the Germans (AFAIK) to 
express a typical access situation on many trunks but also some primaries 
(motorway like access), so that trunk (motorway like physical construction) and 
access could be tagged orthogonally. There are also some trunks that permit 
slower traffic in Germany.

Cheers Martin 


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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Paul Johnson
On Sat, Dec 21, 2019 at 6:37 AM Jarek Piórkowski 
wrote:

> On Fri, 20 Dec 2019 at 22:30, Paul Johnson  wrote:
> >> > What I'm saying is highway=bundesstraße could be acceptable, but
> straße=bundestraße wouldn't be.  Mostly so way type objects with highway=*
> are still potentially routable.
> >>
> >> How do you propose these "potential routable" fallback routers to
> >> handle highway=Bürgersteig (a sidewalk) vs highway=Fahrradstraße (a
> >> local street where bicycles have priority)?
> >
> > Same way I already consider highway=motorway:  Tag for access.
> Assumptions are OKish, but at least in north america, motorway and trunk
> (aka freeways and expressways) are ambiguous for non-motorized modes.  Most
> states allow nonmotorized modes on motorways, but 15 don't, and about 30
> allow it when not otherwise posted, so foot=no, bicycle=no is *NOT* a safe
> assumption for freeways.  Especailly the farther west you go; for example,
> in British Columbia, Washington and Oregon (where I grew up), a motorway is
> probably safer for cyclists than your average city street with bicycle
> lanes (no oncoming traffic, parking on the pavement is not permitted at
> all, roughly 4.5m wide hard shoulders, 80-110km/h speed limits but that's
> 3m from anybody walking or cycling, versus no sidewalks, 1.5-2m wide
> bicycle lanes and 60-90 km/h surface street speed limits, with both the
> speed limits and bicycle lanes being next to never respected in the
> northwest).
>
> Tagging exceptions is fine, but tagging every highway=footway aka
> highway=chodnik with vehicle=no is going to be a bit silly I would
> think.
>
> You might know, but just to mention: the British Columbia comment is
> not accurate in most of Lower Mainland (that one bit of Upper Levels
> Highway notwithstanding). So if we are adding access tags to
> distinguish urban vs rural realities, what's wrong with
> highway=motorway as the base tag?
>
> >> How will a router know
> >> which highways can be used by trucks, buses, pedestrians, other than
> >> with a giant lookup table?
> >
> > highway=* is a good start, which is why I'm in favor of "lower is
> better" with the exiting system.
>
> I'm sorry, I'm missing some context to understand this (or was this
> meant to be "existing"?). To be more explicit, how would a router know
> not to send trucks on highway=Fahrradstraße?
>
> >> And if you do have a giant lookup table,
> >> wouldn't be easier to have it in editors rather than in every single
> >> data consumer?
> >
> > Not really.  UK concepts are entirely foreign to all but roadgeeks in
> North America, and, judging by AARoads and NE2's history, not even then.
>
> Sorry, I meant "editors" as in software, not as in humans. You specify
> the road as "Oregon freeway", and JOSM/rapiD/whatever editor add tags
> highway=motorway + bicycle=yes.
>

Under state law, except where bicycle=no, it appears Oregon is actually
foot=yes, bicycle=designated on all state-operated highways these days,
with the entire state highway system (or it's bicycle bypasses instead
where present) being part of that state's RCN.  Not that this says anything
about actual ridability in Oregon, since most state highways outside of
urban centers do not have bicycle lanes or hard shoulders except for
freeways.  I'm glad that Osmand has "allow motorways" as an option in
bicycle mode, as, if you have the endurance required inherent (emergency
stopping only, even for cyclists, despite potentially extreme distances
between exits; barring physical barriers or mechanical issues preventing
it, you must keep moving) as, especially in western Oregon and the Portland
area specifically, the freeway is probably your best, if not only, bet.

I'd be down for something similar to that.  I'm really not in favor of
using trunk for anything other than expressways though, mostly because
there's an upward creep problem in US tagging already and expressways are a
special case for driving anyway.

Would also be nice if there was an *actually easy* scheme for handling
presets that deal with complicated rules to explicitly tag regional
situations.  Like, for example, you're editing a highway in Oklahoma that
has minspeed=* set, bicycle=no and foot=no automatically get added.  If
that way is part of a bicycle route relation, then it gets
bicycle=designated added automatically, instead.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Mateusz Konieczny



21 Dec 2019, 15:29 by wolfg...@lyxys.ka.sub.org:

> * Mateusz Konieczny 
>
>> 21 Dec 2019, 12:00 by wolfg...@lyxys.ka.sub.org:
>>
>>> I suggest to keep the road classification consistent at least within
>>> a country and try to solve the problem of roads in low-zoom maps at
>>> the rendering level, by modifying the list of displayed road classes
>>> until a target density of displayed roads is reached. That might
>>> become easier to do when we move to vector tiles.
>>>
>> Seems not doable with OSM data - this
>> would require far more road classes
>> than we use.
>>
>
> Why would we need more road classes for that? This would only be an
> issue if the difference between two "adjacent" classes would be so big
> that you would jump from "almost none" to "to many to display" in one
> step.
>
Exactly.

There are many places where motorway 
and trunk is not enough
(as trunks are not used for roads
forming core network but for expressways).

Adding also primary roads pushes it into 
unacceptable many roads forming blobs.

>> lane and surface data is also almost
>> certainly not helpful here even with full
>> coverage
>>
>> And it would result in weird transitions
>> between countries.
>>
>
> Only if road density changes rapidly at the border, and then we would
> just depict the weird transition that exists in reality.
>
In case of using regions not matching countries
you will still have weird transitions on borders of regions.
> I think it might be possible to upgrade the "minimum zoom level to
> display" on a way if there are no already displayed ways in an area,
> maybe only if it connects to an already displayed way (recursive).
> That way we would boost the minimum zoom level of e.g.
> https://www.openstreetmap.org/way/196509120 to zoom-level 11 or maybe
> even 9, even with it being just a low quality dirt track going near an
> obscure archaeological site in the middle of nowhere.
>
I had some ideas, none managed to deal
with "weird borders somewhere".___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Wolfgang Zenker
* Mateusz Konieczny 
> 21 Dec 2019, 12:00 by wolfg...@lyxys.ka.sub.org:
>> I suggest to keep the road classification consistent at least within
>> a country and try to solve the problem of roads in low-zoom maps at
>> the rendering level, by modifying the list of displayed road classes
>> until a target density of displayed roads is reached. That might
>> become easier to do when we move to vector tiles.

> Seems not doable with OSM data - this
> would require far more road classes
> than we use.

Why would we need more road classes for that? This would only be an
issue if the difference between two "adjacent" classes would be so big
that you would jump from "almost none" to "to many to display" in one
step.

> lane and surface data is also almost
> certainly not helpful here even with full
> coverage

> And it would result in weird transitions
> between countries.

Only if road density changes rapidly at the border, and then we would
just depict the weird transition that exists in reality.

I think it might be possible to upgrade the "minimum zoom level to
display" on a way if there are no already displayed ways in an area,
maybe only if it connects to an already displayed way (recursive).
That way we would boost the minimum zoom level of e.g.
https://www.openstreetmap.org/way/196509120 to zoom-level 11 or maybe
even 9, even with it being just a low quality dirt track going near an
obscure archaeological site in the middle of nowhere.

Wolfgang
( lyx@osm )

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Jarek Piórkowski
On Fri, 20 Dec 2019 at 22:30, Paul Johnson  wrote:
>> > What I'm saying is highway=bundesstraße could be acceptable, but 
>> > straße=bundestraße wouldn't be.  Mostly so way type objects with highway=* 
>> >  are still potentially routable.
>>
>> How do you propose these "potential routable" fallback routers to
>> handle highway=Bürgersteig (a sidewalk) vs highway=Fahrradstraße (a
>> local street where bicycles have priority)?
>
> Same way I already consider highway=motorway:  Tag for access.  Assumptions 
> are OKish, but at least in north america, motorway and trunk (aka freeways 
> and expressways) are ambiguous for non-motorized modes.  Most states allow 
> nonmotorized modes on motorways, but 15 don't, and about 30 allow it when not 
> otherwise posted, so foot=no, bicycle=no is *NOT* a safe assumption for 
> freeways.  Especailly the farther west you go; for example, in British 
> Columbia, Washington and Oregon (where I grew up), a motorway is probably 
> safer for cyclists than your average city street with bicycle lanes (no 
> oncoming traffic, parking on the pavement is not permitted at all, roughly 
> 4.5m wide hard shoulders, 80-110km/h speed limits but that's 3m from anybody 
> walking or cycling, versus no sidewalks, 1.5-2m wide bicycle lanes and 60-90 
> km/h surface street speed limits, with both the speed limits and bicycle 
> lanes being next to never respected in the northwest).

Tagging exceptions is fine, but tagging every highway=footway aka
highway=chodnik with vehicle=no is going to be a bit silly I would
think.

You might know, but just to mention: the British Columbia comment is
not accurate in most of Lower Mainland (that one bit of Upper Levels
Highway notwithstanding). So if we are adding access tags to
distinguish urban vs rural realities, what's wrong with
highway=motorway as the base tag?

>> How will a router know
>> which highways can be used by trucks, buses, pedestrians, other than
>> with a giant lookup table?
>
> highway=* is a good start, which is why I'm in favor of "lower is better" 
> with the exiting system.

I'm sorry, I'm missing some context to understand this (or was this
meant to be "existing"?). To be more explicit, how would a router know
not to send trucks on highway=Fahrradstraße?

>> And if you do have a giant lookup table,
>> wouldn't be easier to have it in editors rather than in every single
>> data consumer?
>
> Not really.  UK concepts are entirely foreign to all but roadgeeks in North 
> America, and, judging by AARoads and NE2's history, not even then.

Sorry, I meant "editors" as in software, not as in humans. You specify
the road as "Oregon freeway", and JOSM/rapiD/whatever editor add tags
highway=motorway + bicycle=yes.

I'm totally onboard with additional tags to refine for local
realities. But keep the one tag already understood by all data
consumers.

Thanks,
--Jarek

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


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Mateusz Konieczny



21 Dec 2019, 12:56 by ajt1...@gmail.com:

> On 21/12/2019 11:40, Mateusz Konieczny  wrote:
>
>>
>> 21 Dec 2019, 12:00 by >> wolfg...@lyxys.ka.sub.org>> :
>>
>>> I suggest to keep the road classification consistent at  least 
>>> within
>>> a country and try to solve the problem of roads in low-zoom  maps at
>>> the rendering level, by modifying the list of displayed  road 
>>> classes
>>> until a target density of displayed roads is reached. That  might
>>> become easier to do when we move to vector tiles.
>>>
>> Seems not doable with OSM data - this
>> would require far more road classes
>> than we use.
>>
>> lane and surface data is also almost
>> certainly not helpful here even with full
>> coverage
>>
>
> Renderers can certainly use tags other than "highway" when  deciding what 
> road class to render things as
>
>
I know, that is why I mentioned 
lane and surface tags
>
> Specifically, if a tertiary is particular narrow it gets rendered  as 
> unclassified, and if an unclassified or residential is a gravel  track it 
> gets rendered as a track with public access.
>
>
OSM Carto is using access tag in
this way. It works well for zoomed in
view but gets more problematic for
country level views or global ones.___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Andy Townsend

On 21/12/2019 11:40, Mateusz Konieczny wrote:


21 Dec 2019, 12:00 by wolfg...@lyxys.ka.sub.org:

I suggest to keep the road classification consistent at least within
a country and try to solve the problem of roads in low-zoom maps at
the rendering level, by modifying the list of displayed road classes
until a target density of displayed roads is reached. That might
become easier to do when we move to vector tiles.

Seems not doable with OSM data - this
would require far more road classes
than we use.

lane and surface data is also almost
certainly not helpful here even with full
coverage

Renderers can certainly use tags other than "highway" when deciding what 
road class to render things as (I do that in a couple places in maps for 
UK/IE but not for trunk/primary, but if I was creating maps for the US I 
suspect I'd definitely try and use other tags along with the highway tag 
to influence road class at motorway/trunk/primary as well).


Specifically, if a tertiary is particular narrow it gets rendered as 
unclassified, and if an unclassified or residential is a gravel track it 
gets rendered as a track with public access. Specifically:


https://github.com/SomeoneElseOSM/SomeoneElse-style/blob/master/style.lua#L928

https://github.com/SomeoneElseOSM/SomeoneElse-style/blob/master/style.lua#L421

The "invented highway tags" set there are then used by

https://github.com/SomeoneElseOSM/openstreetmap-carto-AJT/blob/master/roads.mss

Best Regards,

Andy


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


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Mateusz Konieczny

21 Dec 2019, 12:00 by wolfg...@lyxys.ka.sub.org:
> I suggest to keep the road classification consistent at least within
> a country and try to solve the problem of roads in low-zoom maps at
> the rendering level, by modifying the list of displayed road classes
> until a target density of displayed roads is reached. That might
> become easier to do when we move to vector tiles.
>
Seems not doable with OSM data - this
would require far more road classes
than we use.

lane and surface data is also almost
certainly not helpful here even with full
coverage

And it would result in weird transitions
between countries.___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary

2019-12-21 Thread Joseph Eisenberg
> I would expect that a road shown as e.g. trunk in Massachusets would be quite 
> similar in
characteristics to a road shown as trunk in Montana.

Characteristics always change strongly between rural areas and urban
areas: in most places a highway=primary will have several lanes in a
large city, but in a rural area it will usually be only one lane each
way. In a city the speed limit will often been 40 kmh (25 mph), but in
a rural area it might be 80 kmh (55 mph) on the same class of road.
Similarly, near-wilderness areas like Alaska can be expected to have
different road characteristics than more densely-populated rural
areas.

> try to solve the problem of roads in low-zoom maps at the rendering level, by 
> modifying the list of displayed road classes until a target density of 
> displayed roads is reached

That's the right solution for maps that are designed for a particular
region or country, but it is not feasible for automatically generated
maps with global coverage (like the 4 styles shown on
openstreetmap.org). The big problem is what happens at the borders
between areas: it could be quite confusing if the color or zoom level
of a certain highway tag suddenly changed at the border between Spain
and France for example.

- Joseph Eisenberg

On 12/21/19, Wolfgang Zenker  wrote:
> * Joseph Eisenberg  wrote:
>> Above it was said that the highway=trunk vs highway=primary
>> distinction is mostly for routing applications. But allowing a proper
>> rendering is also a main goal of the road tagging system.
>
>> While it's true that road class is useful for routing when there are
>> two alterate routes, a main reason to tag highways with a certain
>> class is to be able to render maps properly at different zoom levels.
>
>> When you are making a high-scale, low-zoom-level map of a large area
>> (say, the whole State of Alaska, all of England, or all of Australia),
>> you will want to only render highway=motorway + highway=trunk, because
>> showing all highway=primary would lead to rendering many smaller roads
>> which are not reasonable to show at that scale in most places.
>
>> [..]
>
> What makes the problem of road classification so hard is that we want it
> to do different things at once. For rendering we have on the one hand
> the requirement that we want to show all the "relevant" roads for a
> given zoom level, on the other hand, as a map user I would expect that
> a road shown as e.g. trunk in Massachusets would be quite similar in
> characteristics to a road shown as trunk in Montana.
>
> I suggest to keep the road classification consistent at least within
> a country and try to solve the problem of roads in low-zoom maps at
> the rendering level, by modifying the list of displayed road classes
> until a target density of displayed roads is reached. That might
> become easier to do when we move to vector tiles.
>
> Wolfgang
> ( lyx@osm )
>
> ___
> 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] [Talk-us] Trunk VS primary

2019-12-21 Thread Wolfgang Zenker
* Joseph Eisenberg  wrote:
> Above it was said that the highway=trunk vs highway=primary
> distinction is mostly for routing applications. But allowing a proper
> rendering is also a main goal of the road tagging system.

> While it's true that road class is useful for routing when there are
> two alterate routes, a main reason to tag highways with a certain
> class is to be able to render maps properly at different zoom levels.

> When you are making a high-scale, low-zoom-level map of a large area
> (say, the whole State of Alaska, all of England, or all of Australia),
> you will want to only render highway=motorway + highway=trunk, because
> showing all highway=primary would lead to rendering many smaller roads
> which are not reasonable to show at that scale in most places.

> [..]

What makes the problem of road classification so hard is that we want it
to do different things at once. For rendering we have on the one hand
the requirement that we want to show all the "relevant" roads for a
given zoom level, on the other hand, as a map user I would expect that
a road shown as e.g. trunk in Massachusets would be quite similar in
characteristics to a road shown as trunk in Montana.

I suggest to keep the road classification consistent at least within
a country and try to solve the problem of roads in low-zoom maps at
the rendering level, by modifying the list of displayed road classes
until a target density of displayed roads is reached. That might
become easier to do when we move to vector tiles.

Wolfgang
( lyx@osm )

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Mateusz Konieczny



21 Dec 2019, 01:09 by joseph.eisenb...@gmail.com:

> Above it was said that the highway=trunk vs highway=primary
> distinction is mostly for routing applications. But allowing a proper
> rendering is also a main goal of the road tagging system.
>
Yes, during my work on road display in
OSM Carto I really wished for consistent
use of highway=trunk as
"The most important roads on national
level, but not motorways".

I encountered some definitions of highway=trunk
describing this type of user,
but it is sadly not universally followed.

I think it would be desirable to use it this way,
rather than for tagging of "high performance
roads below motorway quality".
> While it's true that road class is useful for routing when there are
> two alterate routes, a main reason to tag highways with a certain
> class is to be able to render maps properly at different zoom levels.
>
> When you are making a high-scale, low-zoom-level map of a large area
> (say, the whole State of Alaska, all of England, or all of Australia),
> you will want to only render highway=motorway + highway=trunk, because
> showing all highway=primary would lead to rendering many smaller roads
> which are not reasonable to show at that scale in most places.
>
+1___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-21 Thread Mateusz Konieczny



21 Dec 2019, 01:44 by ba...@ursamundi.org:

>
>
> On Fri, Dec 20, 2019 at 1:07 AM Mateusz Konieczny <> matkoni...@tutanota.com> 
> > wrote:
>
>>
>> 20 Dec 2019, 01:25 by >> ba...@ursamundi.org>> :
>>
>>> So, for example, in the US, instead of motorway, trunk, primary, secondary, 
>>> tertiary, perhaps something more like freeway, expressway, 
>>> major/minor_principal (just having this would fix a *lot* of problems with 
>>> Texas and Missouri and their extensive secondary systems), 
>>> major/minor_collector...the US just has a way more complex view of how 
>>> highways work.  
>>>
>>> Or at least some more serious consideration given to the proposal at >>> 
>>> https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS>>>  (but perhaps 
>>> with "other principal arterials" as primary and a new "highway=quartinary".
>>>
>> Fitting thing like road classification
>> into UK system is irritating at times.
>>
>> But idea of each country with separate tags
>> for roads is simply a bad idea.
>>
>
> Could you expand on this?  Being able to speak each country's highway lingua 
> franca would make it a lot easier for OSM to become the Rosetta Stone of maps 
> simply from ease of classification.
>
I am consider it unlikely that it would make
anything easier.

Current solution is not ideal butfollowing each local and incompatible 
classification scheme instead seems to not be better.

I am 100% OK with tagging official road 
status somehow - US expressway,
US highway route, Polish droga wojewódzka,
Polish droga gminna and so on.

But as a new (maybe already existing)
tag.
But do not expect 1:1 mapping to highway tag value.
>  
>
>>
>> This info is probably worth recording,
>> but legal status should go into a separate tag.
>>
>
> Legal status of roads in the US isn't quite as clearcut as it is in the UK, 
> where the highway=* tag is literally equal to that country's legal 
> classification, plus private roads with significant public passage and/or 
> reach.  Off the top of my head we have 1 country, 2 states, 34 tribes, 77 
> counties and 597 towns, plus MacQuarie Group Australia running the turnpikes 
> and the Boy Scouts of America, Phillips 66, ConocoPhillips, or some 
> combination of the three, and potentially scores more private entities, 
> operating extensive networks of publicly accessible roads and highways in 
> Oklahoma.  And I generally consider myself lucky I have it > this>  
> straightforward in the US.
>
> Texas likely has similar situations but throw in the fact that they have 7 
> different state highway systems before you get into at least 3 more 
> (regional? state? private? unclear...) competing turnpike networks, sometimes 
> running side by side on the same right of way (consider TX 121 with the 
> George Bush Turnpike operated by the North Texas Transportation Agency 
> running down the median).
>
> Simply starting with the HFCS and expanding from that (particularly on the 
> freeway/expressway distinction, and having more levels between secondary and 
> unclassified) would be a fantastic boon to dealing with this mess in a more 
> concise fashion as it changes highway=* tagging from almost entirely 
> subjective to subjective but within a limited range.  Establish wiki pages 
> describing how each region works and let the consumers sort it out from there.
>
> At an absolute minimum, we really need to establish values lower than 
> tertiary yet above unclassified, and we definitely do need to make the 
> freeway/expressway distinction.
>
I consider any plan that would add new
highway values to be unlikely to succeed.

Consider introducing new tags instead.___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 7:47 PM Jarek Piórkowski 
wrote:

> On Fri, 20 Dec 2019 at 20:26, Paul Johnson  wrote:
> >> > I'm not arguing in favor of a change in language for key name.  But
> the local broadly accepted classification terminology (preferably in
> English for consistency sake) for the value.
> >>
> >> Why in English? Bundesstraße is a broadly accepted classification
> >> terminology, so is autostrada. If you want to do things for
> >> consistency sake, there are the accepted OSM-British-English names.
> >
> >
> > What I'm saying is highway=bundesstraße could be acceptable, but
> straße=bundestraße wouldn't be.  Mostly so way type objects with highway=*
> are still potentially routable.
>
> How do you propose these "potential routable" fallback routers to
> handle highway=Bürgersteig (a sidewalk) vs highway=Fahrradstraße (a
> local street where bicycles have priority)?


Same way I already consider highway=motorway:  Tag for access.  Assumptions
are OKish, but at least in north america, motorway and trunk (aka freeways
and expressways) are ambiguous for non-motorized modes.  Most states allow
nonmotorized modes on motorways, but 15 don't, and about 30 allow it when
not otherwise posted, so foot=no, bicycle=no is *NOT* a safe assumption for
freeways.  Especailly the farther west you go; for example, in British
Columbia, Washington and Oregon (where I grew up), a motorway is probably
safer for cyclists than your average city street with bicycle lanes (no
oncoming traffic, parking on the pavement is not permitted at all, roughly
4.5m wide hard shoulders, 80-110km/h speed limits but that's 3m from
anybody walking or cycling, versus no sidewalks, 1.5-2m wide bicycle lanes
and 60-90 km/h surface street speed limits, with both the speed limits and
bicycle lanes being next to never respected in the northwest).


> How will a router know
> which highways can be used by trucks, buses, pedestrians, other than
> with a giant lookup table?


highway=* is a good start, which is why I'm in favor of "lower is better"
with the exiting system.


> And if you do have a giant lookup table,
> wouldn't be easier to have it in editors rather than in every single
> data consumer?
>

Not really.  UK concepts are entirely foreign to all but roadgeeks in North
America, and, judging by AARoads and NE2's history, not even then.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Tod Fitch

> On Dec 20, 2019, at 5:25 PM, Paul Johnson  wrote:
> 
> 
> What I'm saying is highway=bundesstraße could be acceptable, but 
> straße=bundestraße wouldn't be.  Mostly so way type objects with highway=*  
> are still potentially routable.

I sure wouldn’t want to be the person in charge of maintaining either the style 
definition or the SQL select function that had to decide which of many possible 
highway tag values I was going to render as a freeway, major road, minor road, 
etc. when creating a map that covers the whole world or even just a significant 
part of it.

Cheers!
Tod




signature.asc
Description: Message signed with OpenPGP
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Jarek Piórkowski
On Fri, 20 Dec 2019 at 20:26, Paul Johnson  wrote:
>> > I'm not arguing in favor of a change in language for key name.  But the 
>> > local broadly accepted classification terminology (preferably in English 
>> > for consistency sake) for the value.
>>
>> Why in English? Bundesstraße is a broadly accepted classification
>> terminology, so is autostrada. If you want to do things for
>> consistency sake, there are the accepted OSM-British-English names.
>
>
> What I'm saying is highway=bundesstraße could be acceptable, but 
> straße=bundestraße wouldn't be.  Mostly so way type objects with highway=*  
> are still potentially routable.

How do you propose these "potential routable" fallback routers to
handle highway=Bürgersteig (a sidewalk) vs highway=Fahrradstraße (a
local street where bicycles have priority)? How will a router know
which highways can be used by trucks, buses, pedestrians, other than
with a giant lookup table? And if you do have a giant lookup table,
wouldn't be easier to have it in editors rather than in every single
data consumer?

--Jarek

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 7:22 PM Jarek Piórkowski 
wrote:

> On Fri, 20 Dec 2019 at 20:16, Paul Johnson  wrote:
> > On Fri, Dec 20, 2019 at 6:57 PM Joseph Eisenberg <
> joseph.eisenb...@gmail.com> wrote:
> >> > Being able to speak each country's highway lingua franca would make
> it a lot easier for OSM to become the Rosetta Stone of maps simply from
> ease of classification.
> >>
> >> That would mean using "jalan=provinsi" instead of "highway=primary" in
> >> Indonesia, so any global map service (like opencyclemap.org) would
> >> need to interpret all these tags from different languages. If you
> >> limit this to just official languages there would be several hundred
> >> to translate, but there are over 1500 languages with a written
> >> language currently: I don't see why we would limit things to just
> >> official languages.
> >
> >
> > I'm not arguing in favor of a change in language for key name.  But the
> local broadly accepted classification terminology (preferably in English
> for consistency sake) for the value.
>
> Why in English? Bundesstraße is a broadly accepted classification
> terminology, so is autostrada. If you want to do things for
> consistency sake, there are the accepted OSM-British-English names.
>

What I'm saying is highway=bundesstraße could be acceptable, but
straße=bundestraße wouldn't be.  Mostly so way type objects with highway=*
are still potentially routable.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Jarek Piórkowski
On Fri, 20 Dec 2019 at 20:16, Paul Johnson  wrote:
> On Fri, Dec 20, 2019 at 6:57 PM Joseph Eisenberg  
> wrote:
>> > Being able to speak each country's highway lingua franca would make it a 
>> > lot easier for OSM to become the Rosetta Stone of maps simply from ease of 
>> > classification.
>>
>> That would mean using "jalan=provinsi" instead of "highway=primary" in
>> Indonesia, so any global map service (like opencyclemap.org) would
>> need to interpret all these tags from different languages. If you
>> limit this to just official languages there would be several hundred
>> to translate, but there are over 1500 languages with a written
>> language currently: I don't see why we would limit things to just
>> official languages.
>
>
> I'm not arguing in favor of a change in language for key name.  But the local 
> broadly accepted classification terminology (preferably in English for 
> consistency sake) for the value.

Why in English? Bundesstraße is a broadly accepted classification
terminology, so is autostrada. If you want to do things for
consistency sake, there are the accepted OSM-British-English names.

--Jarek

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 6:57 PM Joseph Eisenberg 
wrote:

> > Being able to speak each country's highway lingua franca would make it a
> lot easier for OSM to become the Rosetta Stone of maps simply from ease of
> classification.
>
> That would mean using "jalan=provinsi" instead of "highway=primary" in
> Indonesia, so any global map service (like opencyclemap.org) would
> need to interpret all these tags from different languages. If you
> limit this to just official languages there would be several hundred
> to translate, but there are over 1500 languages with a written
> language currently: I don't see why we would limit things to just
> official languages.
>

I'm not arguing in favor of a change in language for key name.  But the
local broadly accepted classification terminology (preferably in English
for consistency sake) for the value.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 5:22 PM Graeme Fitzpatrick 
wrote:

> While =primary refers to "A major highway linking large towns, in
> developed countries normally with 2 lanes. In areas with worse
> infrastructure road quality may be far worse"
> https://wiki.openstreetmap.org/wiki/Tag:highway%3Dprimary
>

Well, that more or less does it, doesn't it?  Homer, Alaska is almost
exactly a statistically average American city in terms of size, anything
larger than that would be a moderately large city with Juneau, Fairbanks
qualifying as moderately major and Anchorage as quite large.  Keep in mind,
the average US city is only about 5000 people.


> If the only way of getting from A to B is along a particular road, it's
>> very important.
>>
>
> In some of the cases that I'm thinking of, these are the only roads, so
> yes, very important.
>

In which case secondary is going to show up adequately while not
overstating its stature as an even more major road than it is.


> Do we need to broaden our classes of construction/legality somewhat to
>> encompass
>> standards in countries outside the UK?  Probably.  Are the values for the
>> highway key
>> UK-centric and somewhat misleading?  Sure.
>>
>
> Thank you!
>
>
>> Should we map dirt tracks between Hell's Bumhole and Arse-end Of Nowhere
>> as motorways because they're the only
>> way of getting from one place to the other?  No, no, a thousand times no.
>>
>
> & yes, yes, a thousand times yes, :-) I agree that they shouldn't be
> =motorways, BUT, they should be either =trunk or =primary, because, in this
> area, they are the main road, & so are of vital importance
>

But, it's largely been established in the US that highway=trunk is a
special case to mean limited access expressway or controlled access single
carriageway, and primary being either an extremely major locally
significant road or a longer road of national importance.  AK2 might fit
primary for most of its length, and definitely fits trunk for it's
Fairbanks dual-carriageway segment, but I think a stronger case is made
with trunk for the Fairbanks dual carriageway and secondary for the rest.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Joseph Eisenberg
>> This info is probably worth recording,
>> but legal status should go into a separate tag.
>>
> Legal status of roads in the US isn't quite as clearcut as it is in the UK,
> where the highway=* tag is literally equal to that country's legal
> classification, plus private roads with significant public passage and/or
> reach.
> ...
> At an absolute minimum, we really need to establish values lower than
> tertiary yet above unclassified, and we definitely do need to make the
> freeway/expressway distinction.

I agree that adding expressway=yes tags to more roads is a good idea.
It might be useful for rendering and routing in some situations.

If we allow highway=trunk to be used for all major highways (not only
those that have certain physical charateristics) this give us an
additional level to work with: in the UK and Spain, highway=primary
links smaller towns and villages since all the major highways are
highway=trunk, which leaves highway=secondary for pretty minor roads
("B Class" in the UK) linking villages and hamlets, while
highway=tertiary is for quite minor roads to hamlets or small
neighborhoods, and "highway=unclassified" is usually for tiny public
roads connecting farms or individual, isolated houses out in the
countryside

In the USA, most US highways can be highway=trunk or highway=primary,
most State highways are highway=primary or =secondary, and country or
local roads are highway=secondary or highway=tertiary if they are
significant. I use highway=unclassified for very small roads.

I don't know Alaska well, but I suspect that the one-lane, gravel
borough roads should not be highway=unclassified: they would usually
be highway=secondary or highway=tertiary, like most county roads in
the western USA. If they are the main route to the largest small town
within 100 miles, they might be highway=primary.

Here in eastern Indonesia, most of my highway=primary roads have large
sections of gravel, and most highway=secondary are narrow gravel or
dirt roads (though in Java and Bali they will usually be paved).

- Joseph Eisenberg

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Joseph Eisenberg
> Being able to speak each country's highway lingua franca would make it a lot 
> easier for OSM to become the Rosetta Stone of maps simply from ease of 
> classification.

That would mean using "jalan=provinsi" instead of "highway=primary" in
Indonesia, so any global map service (like opencyclemap.org) would
need to interpret all these tags from different languages. If you
limit this to just official languages there would be several hundred
to translate, but there are over 1500 languages with a written
language currently: I don't see why we would limit things to just
official languages.

The main feature tags are in British English and they should be
translated to the appropriate local context by local mappers in each
area, rather than creating new feature tags for every country and
language, so that global maps and routing applications can continue to
work.

It's also helpful that mappers in Germany and Japan can help map my
area here in Indonesia, adding rivers, lakes and roads based on aerial
imagery. They would have trouble if they needed to learn the hundreds
of local languages in each part of Indonesia to tag things properly.

-Joseph Eisenberg

On 12/21/19, Paul Johnson  wrote:
> On Fri, Dec 20, 2019 at 1:07 AM Mateusz Konieczny 
> wrote:
>
>>
>> 20 Dec 2019, 01:25 by ba...@ursamundi.org:
>>
>> So, for example, in the US, instead of motorway, trunk, primary,
>> secondary, tertiary, perhaps something more like freeway, expressway,
>> major/minor_principal (just having this would fix a *lot* of problems
>> with
>> Texas and Missouri and their extensive secondary systems),
>> major/minor_collector...the US just has a way more complex view of how
>> highways work.
>>
>> Or at least some more serious consideration given to the proposal at
>> https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS (but perhaps
>> with "other principal arterials" as primary and a new
>> "highway=quartinary".
>>
>> Fitting thing like road classification
>> into UK system is irritating at times.
>>
>> But idea of each country with separate tags
>> for roads is simply a bad idea.
>>
>
> Could you expand on this?  Being able to speak each country's highway
> lingua franca would make it a lot easier for OSM to become the Rosetta
> Stone of maps simply from ease of classification.
>
>
>> This info is probably worth recording,
>> but legal status should go into a separate tag.
>>
>
> Legal status of roads in the US isn't quite as clearcut as it is in the UK,
> where the highway=* tag is literally equal to that country's legal
> classification, plus private roads with significant public passage and/or
> reach.  Off the top of my head we have 1 country, 2 states, 34 tribes, 77
> counties and 597 towns, plus MacQuarie Group Australia running the
> turnpikes and the Boy Scouts of America, Phillips 66, ConocoPhillips, or
> some combination of the three, and potentially scores more private
> entities, operating extensive networks of publicly accessible roads and
> highways in Oklahoma.  And I generally consider myself lucky I have it
> *this* straightforward in the US.
>
> Texas likely has similar situations but throw in the fact that they have 7
> different state highway systems before you get into at least 3 more
> (regional? state? private? unclear...) competing turnpike networks,
> sometimes running side by side on the same right of way (consider TX 121
> with the George Bush Turnpike operated by the North Texas Transportation
> Agency running down the median).
>
> Simply starting with the HFCS and expanding from that (particularly on the
> freeway/expressway distinction, and having more levels between secondary
> and unclassified) would be a fantastic boon to dealing with this mess in a
> more concise fashion as it changes highway=* tagging from almost entirely
> subjective to subjective but within a limited range.  Establish wiki pages
> describing how each region works and let the consumers sort it out from
> there.
>
> At an absolute minimum, we really need to establish values lower than
> tertiary yet above unclassified, and we definitely do need to make the
> freeway/expressway distinction.
>

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 4:41 PM Mateusz Konieczny 
wrote:

>
>
>
> 20 Dec 2019, 23:04 by graemefi...@gmail.com:
>
>
>
>
> On Fri, 20 Dec 2019 at 19:18, Martin Koppenhoefer 
> wrote:
>
>
>
> > On 20. Dec 2019, at 04:02, Graeme Fitzpatrick 
> wrote:
> >
> > that [/the/] (one & only) road servicing an area is dirt, if you're
> lucky, 2 lanes wide, but is used constantly by heavy traffic (semi-trailers
> with 3 o4 4 trailers on the back).
> >
> > How do we tie that into the nice neat Motorway > Primary > Secondary etc
> arrangement?
>
>
> lanes=2
> surface=unpaved
>
>
> Thanks, Martin :-)
>
> But would they still count as either =trunk or =primary?
>
> While they're of high local importance, they're definitely not
> high-performance & they don't link major population centres either?
>
> Main road linking villages (even unpaved
> and unmaintained) is
> highway=unclassified (yes, confusing name
> is confusing).
>

The British idea of "unclassified" maps nicely in Alaska to borough roads
(I would generally assume to be paved or at least graded if not graded and
graveled), since the usable space on U roads in the UK is substantially
smaller than tertiary and lanes are typically not present, or even singular
despite two way travel.


> Maybe this fits here?
>

Eeeh, maybe, but a long stretch.  In the US, it's generally established
that state highways are highway=secondary, even if unpaved, in the lower 48
and Hawaii.  Though the "unpaved" part has been obsolete since AR 220 went
from surface=dirt (it wasn't even graveled) to surface=asphalt, 4 years ago
in two weeks in the lower 48 and Hawaii.  Alaska is currently the only
state, territory or posession of the US that has unpaved state highways and
the last of the same that will ever have unpaved highways in the Eisenhower
Interstate Defense Highway System (despite having no signed network US:I
routes presently).  With that context, AK 2 fits pretty obviously as
highway=secondary.  Based on my look just now, looks like that's a good fit
saved for portions within Fairbanks metro, which would be a good
highway=trunk segment.  There are no portions that I see as rising to
highway=motorway, despite being presently mapped as such in some parts of
Fairbanks.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Johnson
On Fri, Dec 20, 2019 at 1:07 AM Mateusz Konieczny 
wrote:

>
> 20 Dec 2019, 01:25 by ba...@ursamundi.org:
>
> So, for example, in the US, instead of motorway, trunk, primary,
> secondary, tertiary, perhaps something more like freeway, expressway,
> major/minor_principal (just having this would fix a *lot* of problems with
> Texas and Missouri and their extensive secondary systems),
> major/minor_collector...the US just has a way more complex view of how
> highways work.
>
> Or at least some more serious consideration given to the proposal at
> https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS (but perhaps
> with "other principal arterials" as primary and a new "highway=quartinary".
>
> Fitting thing like road classification
> into UK system is irritating at times.
>
> But idea of each country with separate tags
> for roads is simply a bad idea.
>

Could you expand on this?  Being able to speak each country's highway
lingua franca would make it a lot easier for OSM to become the Rosetta
Stone of maps simply from ease of classification.


> This info is probably worth recording,
> but legal status should go into a separate tag.
>

Legal status of roads in the US isn't quite as clearcut as it is in the UK,
where the highway=* tag is literally equal to that country's legal
classification, plus private roads with significant public passage and/or
reach.  Off the top of my head we have 1 country, 2 states, 34 tribes, 77
counties and 597 towns, plus MacQuarie Group Australia running the
turnpikes and the Boy Scouts of America, Phillips 66, ConocoPhillips, or
some combination of the three, and potentially scores more private
entities, operating extensive networks of publicly accessible roads and
highways in Oklahoma.  And I generally consider myself lucky I have it
*this* straightforward in the US.

Texas likely has similar situations but throw in the fact that they have 7
different state highway systems before you get into at least 3 more
(regional? state? private? unclear...) competing turnpike networks,
sometimes running side by side on the same right of way (consider TX 121
with the George Bush Turnpike operated by the North Texas Transportation
Agency running down the median).

Simply starting with the HFCS and expanding from that (particularly on the
freeway/expressway distinction, and having more levels between secondary
and unclassified) would be a fantastic boon to dealing with this mess in a
more concise fashion as it changes highway=* tagging from almost entirely
subjective to subjective but within a limited range.  Establish wiki pages
describing how each region works and let the consumers sort it out from
there.

At an absolute minimum, we really need to establish values lower than
tertiary yet above unclassified, and we definitely do need to make the
freeway/expressway distinction.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Allen
On Fri, 20 Dec 2019 at 23:23, Graeme Fitzpatrick 
wrote:

>
> Thanks, Paul - I don't disagree with a word you said, except maybe the
> importance of road construction?
>

It's complicated, but you understand the difference between a motorway (or
whatever
it is called where you are) and a dirt track.

>
> I mentioned "performance" & "importance" as they are the definitions used
> for highway=trunk
> https://wiki.openstreetmap.org/wiki/Tag:highway%3Dtrunk
>

Doesn't mean those are universally applicalbe.  Those are largely UK
definitions.
They are basically routeing definitions.  An A-road is a good way of
getting from
A to B and a B-road is a worse way of doing it,  Depending upon whether your
metric is shortest distance or best road surface or fastest speed if there
is no
other traffic or,,,

>
> In some of the cases that I'm thinking of, these are the only roads, so
> yes, very important.
>
>
But that's still a routeing question.  Yeah, it's the only way from A to
B.  But it's
still  a dirt track, not a motorway.  Use the standard query tool to ask
how to
get from A to B and it will find it.

 we map dirt tracks between Hell's Bumhole and Arse-end Of Nowhere as
> motorways because they're the only way of getting from one place to the
> other?  No, no, a
>
thousand times no.
>
> & yes, yes, a thousand times yes, :-) I agree that they shouldn't be
> =motorways, BUT, they should be either =trunk or =primary, because, in this
> area, they are the main road, & so are of vital importance!
>

Nope, it's down to construction/legality.  Your "vital/importance"
distinction is mere
routeing.  We in the UK have blurred the distinction between
construction/legality
issues and routeing.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Joseph Eisenberg
Above it was said that the highway=trunk vs highway=primary
distinction is mostly for routing applications. But allowing a proper
rendering is also a main goal of the road tagging system.

While it's true that road class is useful for routing when there are
two alterate routes, a main reason to tag highways with a certain
class is to be able to render maps properly at different zoom levels.

When you are making a high-scale, low-zoom-level map of a large area
(say, the whole State of Alaska, all of England, or all of Australia),
you will want to only render highway=motorway + highway=trunk, because
showing all highway=primary would lead to rendering many smaller roads
which are not reasonable to show at that scale in most places.

In England, where these tags were developed, the distinction between
highway=trunk and highway=primary is subtle: both are "A" roads in the
official classification system, but highway=trunk has a special
sub-classification which says they are more important than other "A"
roads (tagged as primary): "UK OSM users follow the practice that all
green-signed A routes (ie primary routes) are tagged highway=trunk,
while black-and-white-signed A-roads (ie non-primary routes) are
tagged highway=primary".

Thus in the USA it's reasonable to use highway=primary for most State
and some US highways, while the most significant ones which connect
cities and large towns would be tagged highway=trunk.

Look at England at z7 on the Openstreetmap-carto style (the highest
level where highway=primary is not shown):
https://www.openstreetmap.org/#map=8/52.017/-0.261

The highway=trunk roads shown here are the main routes between cities
and towns. Zooming in to z8 shows a dense network of highway=primary
roads connecting smaller towns and large villages to towns and cities,
which would not be reasonable to show at z7

Unfortunately, the road classification system in parts of Continental
Europe was different, so mappers in some major countries, including
Germany and France, chose to use highway=trunk as synonym for
"motorroad" (somewhat similar to a U.S.A. "expressway"), with other
major roads tagged as highway=primary. If you look around the
Openstreetmap-carto rendering of Europ at z7, you will see many gaps
in the rendered road network in these countries and surrounding areas
that use the same system.

Compare Spain and Romania, which instead use highway=trunk for all major non-
motorway roads between cities: here the country-wide road network is
clearly visible with showing just highway=trunk and highway=motorway
at z6 and z7.

In the USA, it's fine to limit highway=trunk to expressways in eastern
States where all the important US highways are expressways and these
form a dense network connecting all cities and towns. But in
sparsely-populated Western states even some of the Interstate highways
are not fully motorways, and almost all US highways are just 2 lanes
(one each way) in the area between the Cascades and the Rocky
Mountains, even those that are the main cross-State routes. If we
don't tag these highways as highway=trunk it isn't possible to render
this area in a reasonable way while using the same rendering rules for
the whole USA.

Major US and State highways between cities, like AK-2 and CA-199,
CA-299, US 97 (main route in Eastern Oregon) and US 101 should be
tagged as more significant than a tiny State road in Delaware which
only connects small towns and villages.

I would suggest looking at the Indonesian road tagging guidelines
(which I was not involved in developing, but I use in mapping
locally): they show very different road quality between the developed
areas and the remote parts of the country:
https://wiki.openstreetmap.org/wiki/Indonesian_Tagging_Guidelines#Examples_of_road_class_in_several_provinces_in_Indonesia.
Most trunk roads are only 1 lane each way, but they are still the
main, National road connecting the large cities on each island. This
should be expected in other large countries like the USA, Australia
and Canada.

For tagging the status of a road as a "motorroad" or "expressway" I
would recommend using the tags motorroad=yes and expressway=yes,
rather than tagging all expressways and motorroads as highway=trunk no
matter their classification or significance in the road network. And
adding maxspeed=, surface=, lanes= and access=  will allow routing
applications and specialized renderers to treat these roads properly.

Joseph Eisenberg

On 12/21/19, Graeme Fitzpatrick  wrote:
> On Sat, 21 Dec 2019 at 08:53, Paul Allen  wrote:
>
>> On Fri, 20 Dec 2019 at 22:05, Graeme Fitzpatrick 
>> wrote:
>>
>>>
>>> But would they still count as either =trunk or =primary?
>>>
>>> While they're of high local importance, they're definitely not
>>> high-performance & they don't link major population centres either?
>>>
>>
>> You have just identified three orthogonal dimensions:
>>
>>- Construction (what you call "performance": motorway or dirt track)
>>- Traffic (number of 

Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Martin Koppenhoefer


sent from a phone

> On 20. Dec 2019, at 23:05, Graeme Fitzpatrick  wrote:
> 
>> lanes=2
>> surface=unpaved
> 
> Thanks, Martin :-)
> 
> But would they still count as either =trunk or =primary?
> 
> While they're of high local importance, they're definitely not 
> high-performance & they don't link major population centres either?


secondary? 

If they are sufficiently important (for a big area, which maybe doesn’t have 
major population centres), I would consider primary. In Germany or Italy roads 
like this would not qualify for trunk, because we’re using the tag only for 
roads without level grade intersections, but AFAIK there is no globally agreed 
upon definition for trunk.


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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Graeme Fitzpatrick
On Sat, 21 Dec 2019 at 08:53, Paul Allen  wrote:

> On Fri, 20 Dec 2019 at 22:05, Graeme Fitzpatrick 
> wrote:
>
>>
>> But would they still count as either =trunk or =primary?
>>
>> While they're of high local importance, they're definitely not
>> high-performance & they don't link major population centres either?
>>
>
> You have just identified three orthogonal dimensions:
>
>- Construction (what you call "performance": motorway or dirt track)
>- Traffic (number of vehicles per hour)
>- "Importance" (read on)
>
> Thanks, Paul - I don't disagree with a word you said, except maybe the
importance of road construction?

I mentioned "performance" & "importance" as they are the definitions used
for highway=trunk
https://wiki.openstreetmap.org/wiki/Tag:highway%3Dtrunk

"Use highway =trunk for
high performance or high importance roads that don't meet the requirement
for motorway .
In different countries, either performance or importance is used as the
defining criterion for trunk"

While =primary refers to "A major highway linking large towns, in developed
countries normally with 2 lanes. In areas with worse infrastructure road
quality may be far worse"
https://wiki.openstreetmap.org/wiki/Tag:highway%3Dprimary


> If the only way of getting from A to B is along a particular road, it's
> very important.
>

In some of the cases that I'm thinking of, these are the only roads, so
yes, very important.


> Do we need to broaden our classes of construction/legality somewhat to
> encompass
> standards in countries outside the UK?  Probably.  Are the values for the
> highway key
> UK-centric and somewhat misleading?  Sure.
>

Thank you!


> Should we map dirt tracks between Hell's Bumhole and Arse-end Of Nowhere
> as motorways because they're the only
> way of getting from one place to the other?  No, no, a thousand times no.
>

& yes, yes, a thousand times yes, :-) I agree that they shouldn't be
=motorways, BUT, they should be either =trunk or =primary, because, in this
area, they are the main road, & so are of vital importance!

  Thanks

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Paul Allen
On Fri, 20 Dec 2019 at 22:05, Graeme Fitzpatrick 
wrote:

>
> But would they still count as either =trunk or =primary?
>
> While they're of high local importance, they're definitely not
> high-performance & they don't link major population centres either?
>

You have just identified three orthogonal dimensions:

   - Construction (what you call "performance": motorway or dirt track)
   - Traffic (number of vehicles per hour)
   - "Importance" (read on)

Depending on the country (or area of a country) those three may be highly
correlated or loosely correlated.  And even within a given area of a given
country, the degree of correlation may vary.

I put "importance" in quotes because I think it's really routeing.  If the
only
way of getting from A to B is along a particular road, it's very
important.  If
you can get from A to B via many routes, but one is by far the shortest,
it's
probably important.  OTOH if, of the many routes from A to B, one is very
much shorter than the others but one of the others takes far less time to
traverse then that second one is probably more important.  So I think
"importance" really means "routeing" (feel free to contradict me).

So how do we represent these three dimensions on a map?  I'd say we go
primarily by construction (with some concession to legalities, because
motorways/
freeways/autobahns have legal constraints that differ from other highways).
  Traffic
rates are irrelevant to what is physically there (but have some bearing on
choosing a
route) and routeing is a decision left to algorithms.

Note that routeing algorithms can be run by a computer or they can be run
by wetware in a human brain looking at representations on a map.  So...

Is the route between A and B a motorway or a dirt track?  This is important
for
routeing algorithms on computers and in human wetware because it indicates
typical speed limits, legal requirements, etc.  Is this the shortest route
between
A and B?  A routeing algorithm can determine that whether it's represented
as
a motorway or a dirt track, and so can a human looking at a map.  Is it the
fastest route between A and B?  A routeing algorithm can determine that from
speed limits, but wetware is aided by seeing if the map renders it as a
motorway
or a dirt track.  Is it important?  Depends whether you want to go from A
to B or not,
but the desirability of using it to get from A to B is down to routeing
algorithms
(whether computer or human wetware).

So I come down firmly on construction (as modified by legislation0.  That's
verifiable.
All else is essentially routeing, whether done by computer or done by
wetware in a
human brain looking at a map.  Want to get from A to B?  That's what the
query tool
in standard carto is for.  Maybe one day it will offer the choice of
"fastest" and
"shortest" (maybe it already does).  Or just look at the map, as we boomers
are
used to from looking at printed maps, and figure out what the best route
might be.

Do we need to broaden our classes of construction/legality somewhat to
encompass
standards in countries outside the UK?  Probably.  Are the values for the
highway key
UK-centric and somewhat misleading?  Sure.  Should we map dirt tracks
between
Hell's Bumhole and Arse-end Of Nowhere as motorways because they're the only
way of getting from one place to the other?  No, no, a thousand times no.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Mateusz Konieczny



20 Dec 2019, 23:04 by graemefi...@gmail.com:

>
>
>
> On Fri, 20 Dec 2019 at 19:18, Martin Koppenhoefer <> dieterdre...@gmail.com> 
> > wrote:
>
>>
>>
>> > On 20. Dec 2019, at 04:02, Graeme Fitzpatrick <>> graemefi...@gmail.com>> 
>> > > wrote:
>>  > 
>>  > that [/the/] (one & only) road servicing an area is dirt, if you're 
>> lucky, 2 lanes wide, but is used constantly by heavy traffic (semi-trailers 
>> with 3 o4 4 trailers on the back).
>>  > 
>>  > How do we tie that into the nice neat Motorway > Primary > Secondary etc 
>> arrangement?
>>  
>>  
>>  lanes=2
>>  surface=unpaved
>>
>
> Thanks, Martin :-)
>
> But would they still count as either =trunk or =primary?
>
> While they're of high local importance, they're definitely not 
> high-performance & they don't link major population centres either?
>
Main road linking villages (even unpaved
and unmaintained) is 
highway=unclassified (yes, confusing name
is confusing).

Maybe this fits here?

There are also secondary, tertiary
for main links between 
smaller settlements (and also the main
rotate within cities).___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Graeme Fitzpatrick
On Fri, 20 Dec 2019 at 19:18, Martin Koppenhoefer 
wrote:

>
>
> > On 20. Dec 2019, at 04:02, Graeme Fitzpatrick 
> wrote:
> >
> > that [/the/] (one & only) road servicing an area is dirt, if you're
> lucky, 2 lanes wide, but is used constantly by heavy traffic (semi-trailers
> with 3 o4 4 trailers on the back).
> >
> > How do we tie that into the nice neat Motorway > Primary > Secondary etc
> arrangement?
>
>
> lanes=2
> surface=unpaved
>

Thanks, Martin :-)

But would they still count as either =trunk or =primary?

While they're of high local importance, they're definitely not
high-performance & they don't link major population centres either?

  Thanks

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-20 Thread Martin Koppenhoefer


sent from a phone

> On 20. Dec 2019, at 04:02, Graeme Fitzpatrick  wrote:
> 
> that [/the/] (one & only) road servicing an area is dirt, if you're lucky, 2 
> lanes wide, but is used constantly by heavy traffic (semi-trailers with 3 o4 
> 4 trailers on the back).
> 
> How do we tie that into the nice neat Motorway > Primary > Secondary etc 
> arrangement?


lanes=2
surface=unpaved


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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-19 Thread Phake Nick
Problem with applying different road classification system from different
places with their individual tags onto local roads is that:
1. Even if we ignore countries that have different rules within different
part of a single country, there are still about 200 countries in this
world. Each of them having different road classification values, supposedly
each country should have at least a few different levels, would mean there
are thousand of different values that data consumer or renderer or routing
engine and such need to handle and process, which is a pretty big task for
anyone who wish to use these data. It also make inter-country data
comparison difficult or impossible.
2. It will be difficult for a mapper from country A to map for country B
which might be less mapped through means like satellite mapping, video
mapping or site visit, as the mapper would need to familiar themselves with
local classification instead of applying universal standards onto them.
3. Road classification system within each single country is not always
static. Country sovereignty on localities are not static either. When
either of that change, you would need to re-tag every single road in that
place or that country to the new standard if you decided to introduce
country-specific tagging, which would be a very big task.
4. If you are going to setup a table to tell people which road level in
which country is equal to what in other countries, then why not just tell
people with a table that these roads which are tagged with universal tags
corresponds to those level in local classification?
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-19 Thread Mateusz Konieczny



20 Dec 2019, 01:25 by ba...@ursamundi.org:

> So, for example, in the US, instead of motorway, trunk, primary, secondary, 
> tertiary, perhaps something more like freeway, expressway, 
> major/minor_principal (just having this would fix a *lot* of problems with 
> Texas and Missouri and their extensive secondary systems), 
> major/minor_collector...the US just has a way more complex view of how 
> highways work.  
>
> Or at least some more serious consideration given to the proposal at > 
> https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS>  (but perhaps 
> with "other principal arterials" as primary and a new "highway=quartinary".
>
Fitting thing like road classification
into UK system is irritating at times.

But idea of each country with separate tags
for roads is simply a bad idea.

This info is probably worth recording,
but legal status should go into a separate tag.

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


Re: [Tagging] [Talk-us] Trunk VS primary,

2019-12-19 Thread Paul Johnson
On Thu, Dec 19, 2019 at 1:19 PM Martijn van Exel  wrote:

> I actually like your suggestion that highway=trunk does not add much value
> to the U.S. map, Eric.
> We love to add detail / granularity to OSM so much, it can become hard to
> envisage taking some away.
> Not saying we should abolish trunk right here and now, but something I'd
> consider as one outcome.
>

I'd like to see a lot more left up to the data consumer and more regional
values to be widely acceptable.  For example, instead of trying to smash
the entire planet into the UK's prescribed values and trying to come up
with equivalences, use the terminology each country uses.  So, for example,
in the US, instead of motorway, trunk, primary, secondary, tertiary,
perhaps something more like freeway, expressway, major/minor_principal
(just having this would fix a *lot* of problems with Texas and Missouri and
their extensive secondary systems), major/minor_collector...the US just has
a way more complex view of how highways work.

Or at least some more serious consideration given to the proposal at
https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS (but perhaps
with "other principal arterials" as primary and a new "highway=quartinary".

Much like moving route refs to highway relations (freeing the ref=* tag on
highways for situations where the road and the route have different refs),
leaving the mental gymnastics up to an algorithm and leaving less confusion
to the mapper is getting to be long overdue.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging