Re: [Talk-us] US highway classification (trunk)

[Nathan Edgars II]

On 6/2/2011 3:17 PM, Nathan Edgars II wrote:

To this end, I've been systematically going through trunks in
the US and adding lanes=* tags. This is of course useful even if nothing
is done rendering-wise.


Thanks to PeterIto, we can see the fruits of this: 
http://www.itoworld.com/product/data/ito_map/main?view=128&lat=38&lon=-97&zoom=6

red=2 lanes, green=4 lanes, blue=6 lanes...
Trunk and motorway are wider than primary, so thick red are two-lane 
trunks (or motorways where they exist).


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

Re: [Talk-us] US highway classification (trunk)

[Nathan Edgars II]

On 5/29/2011 3:32 AM, Nathan Mills wrote:

On Sun, 29 May 2011 03:00:03 -0400, Nathan Edgars II wrote:


Perhaps the best way to handle it would be to render a wider line if
oneway=yes and not lanes=1 or if oneway=no/unset and lanes=4 or more.
Thus divided highways would not need a lane count to be wider, but
undivided roads would need to be tagged as having four lanes.


That seems like it would be a reasonable way to handle it. I'm no mapnik
expert, but I can see what I can figure out with the installation I have
locally.


Thank you. To this end, I've been systematically going through trunks in 
the US and adding lanes=* tags. This is of course useful even if nothing 
is done rendering-wise.


By the way, if anyone wants to do some armchair mapping and needs 
suggestions, http://jxapi.openstreetmap.org/xapi/api/0.6/*[FIXME=dual 
carriageway] is a good starting point. Just make sure to correctly 
modify the lane count, relation roles, and anything else.


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

Re: [Talk-us] US highway classification

[Kristian Zoerhoff]

On Tue, May 31, 2011 at 6:21 PM, Paul Johnson  wrote:
> On 05/31/2011 06:26 AM, Kristian Zoerhoff wrote:
>> On Mon, May 30, 2011 at 10:41 PM, Toby Murray 
>>  wrote:
>>> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills 
>>>  wrote:
 On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:

> I'm thinking the differences between motorways and trunks are minor.
> Trunks may have intersections, motorways don't.

 That's the simple way to state my opinion. It also seemed to be the thrust
 of most of the discussion on the talk page of the wiki page referenced
 previously as closest to consensus (the page itself just references the
 existence of the two camps and leaves it at that).

 In short, my position is simply that an end user expects a trunk road to be
 identifiably different than primary or secondary. That's how it's done on
 other maps, so I don't see why that's such a bad thing here.
>>>
>>> I agree with this as well. And I too thought this was a pretty widely
>>> accepted convention.
>>
>> That's one accepted convention, to be sure, but it sometimes ignores
>> the realities of where traffic goes.
>>
>> To give an example: 
>
> 59 and 19...which networks?  Those two routes have incomplete refs.

IL state routes. I'll get to them someday, maybe.

>> If we stuck purely to the above
>> convention, 72 would be trunk, and 20 would be primary (at best).  But
>> traffic flow cares more about where the road goes, not what it looks
>> like.
>
> I'd probably consider both 20 and 72 as trunks based on their design
> looking at the NAIP footage.

Maybe. I definitely think IL 72 should be primary further west than it
is today, but I got sidetracked into fixing more bad TIGER alignments.

-- 
Kristian Zoerhoff
kristian.zoerh...@gmail.com

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

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/31/2011 06:26 AM, Kristian Zoerhoff wrote:
> On Mon, May 30, 2011 at 10:41 PM, Toby Murray 
>  wrote:
>> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills 
>>  wrote:
>>> On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:
>>>
 I'm thinking the differences between motorways and trunks are minor.
 Trunks may have intersections, motorways don't.
>>>
>>> That's the simple way to state my opinion. It also seemed to be the thrust
>>> of most of the discussion on the talk page of the wiki page referenced
>>> previously as closest to consensus (the page itself just references the
>>> existence of the two camps and leaves it at that).
>>>
>>> In short, my position is simply that an end user expects a trunk road to be
>>> identifiably different than primary or secondary. That's how it's done on
>>> other maps, so I don't see why that's such a bad thing here.
>>
>> I agree with this as well. And I too thought this was a pretty widely
>> accepted convention.
> 
> That's one accepted convention, to be sure, but it sometimes ignores
> the realities of where traffic goes.
> 
> To give an example: 

59 and 19...which networks?  Those two routes have incomplete refs.

> If we stuck purely to the above
> convention, 72 would be trunk, and 20 would be primary (at best).  But
> traffic flow cares more about where the road goes, not what it looks
> like.

I'd probably consider both 20 and 72 as trunks based on their design
looking at the NAIP footage.



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Dale Puch]

I have only skimmed these messages, so forgive me if it was already brought
up.
There are two criteria I do not think were brought up.  Length of a road, ie
is it important for the city, county, state, or country.  This needs to be
balanced with the width, and other features of the road like intersections
ect.
The other is relative importance of the road.  I know this is subjective,
but for places without many roads, even a dirt road might be a main
connector between points.

In the end this is a map, and it needs to inform of the best roads to get
from place to place.  This will depend on the map scale and distance
traveled.  Longer roads, especially ones with good throughput should
generally be the higher class roads.

Dale

On Tue, May 31, 2011 at 10:37 AM, Kristian Zoerhoff <
kristian.zoerh...@gmail.com> wrote:

> I hate it when I forget to hit Reply-All
>
> On Tue, May 31, 2011 at 8:54 AM, Greg Troxel  wrote:
> >
> > Kristian Zoerhoff  writes:
> >
> >> On Mon, May 30, 2011 at 10:41 PM, Toby Murray 
> wrote:
> >>> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills 
> wrote:
>  On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:
> 
> > I'm thinking the differences between motorways and trunks are minor.
> > Trunks may have intersections, motorways don't.
> 
>  That's the simple way to state my opinion. It also seemed to be the
> thrust
>  of most of the discussion on the talk page of the wiki page referenced
>  previously as closest to consensus (the page itself just references
> the
>  existence of the two camps and leaves it at that).
> 
>  In short, my position is simply that an end user expects a trunk road
> to be
>  identifiably different than primary or secondary. That's how it's done
> on
>  other maps, so I don't see why that's such a bad thing here.
> >>>
> >>> I agree with this as well. And I too thought this was a pretty widely
> >>> accepted convention.
> >>
> >> That's one accepted convention, to be sure, but it sometimes ignores
> >> the realities of where traffic goes.
> >>
> >> To give an example: 
> >>
> >> IL 72 (the secondary at the top of the map) is a 4- to 6-lane at-grade
> >> expressway; wide median, lights only every mile or so, speed limit up
> >> to 55 mph. It carries a fair amount of traffic, but because it
> >> parallels I 90 (a toll road here), it really only peaks at rush hour,
> >> when the toll road is near capacity..
> >>
> >> US 20 (the trunk at the map bottom), is a 4-lane, non-divided road,
> >> but it carries far more traffic than 72, as it connects the two
> >> motorways at the map ends (the Elgin-O'Hare Expressway, and the Elgin
> >> Bypass, which were never connected). It's not particularly
> >> distinguishable from a lesser 4-lane road, aside from the absurd
> >> amount of traffic it carries. If we stuck purely to the above
> >> convention, 72 would be trunk, and 20 would be primary (at best).  But
> >
> > But what's wrong with that?  It sounds like IL 72 is a higher-class road
> > in terms of the physical road, and US 20 doesn't seem to have
> > almost-motorway features.   Just because a road that is properly
> > labeled primary is heavily used doesn't make it a higher class; you
> > certainly wouldn't label it a motorway based on traffic count.
>
> No, but motorways are such a special case of highway I really don't
> think we should use them as a basis of comparison. You're either a
> motorway, or you aren't.
>
> >> traffic flow cares more about where the road goes, not what it looks
> >> like.
> >
> > Sure, and routers can use that.
> >
> >
> > Probably we need to completely decouple
> >
> >  nominal importance in the hierarchy of road types
> >  physical characteristics
> >  importance to the people who use it
>
> Haven't we already? Physical characteristics have tags (surface,
> lanes, maxspeed). It's the hierarchy that seems to be the sticking
> point, and that's exactly what I thought "classification" was.
>
> --
> Kristian Zoerhoff
> kristian.zoerh...@gmail.com
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Kristian Zoerhoff]

I hate it when I forget to hit Reply-All

On Tue, May 31, 2011 at 8:54 AM, Greg Troxel  wrote:
>
> Kristian Zoerhoff  writes:
>
>> On Mon, May 30, 2011 at 10:41 PM, Toby Murray  wrote:
>>> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills  wrote:
 On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:

> I'm thinking the differences between motorways and trunks are minor.
> Trunks may have intersections, motorways don't.

 That's the simple way to state my opinion. It also seemed to be the thrust
 of most of the discussion on the talk page of the wiki page referenced
 previously as closest to consensus (the page itself just references the
 existence of the two camps and leaves it at that).

 In short, my position is simply that an end user expects a trunk road to be
 identifiably different than primary or secondary. That's how it's done on
 other maps, so I don't see why that's such a bad thing here.
>>>
>>> I agree with this as well. And I too thought this was a pretty widely
>>> accepted convention.
>>
>> That's one accepted convention, to be sure, but it sometimes ignores
>> the realities of where traffic goes.
>>
>> To give an example: 
>>
>> IL 72 (the secondary at the top of the map) is a 4- to 6-lane at-grade
>> expressway; wide median, lights only every mile or so, speed limit up
>> to 55 mph. It carries a fair amount of traffic, but because it
>> parallels I 90 (a toll road here), it really only peaks at rush hour,
>> when the toll road is near capacity..
>>
>> US 20 (the trunk at the map bottom), is a 4-lane, non-divided road,
>> but it carries far more traffic than 72, as it connects the two
>> motorways at the map ends (the Elgin-O'Hare Expressway, and the Elgin
>> Bypass, which were never connected). It's not particularly
>> distinguishable from a lesser 4-lane road, aside from the absurd
>> amount of traffic it carries. If we stuck purely to the above
>> convention, 72 would be trunk, and 20 would be primary (at best).  But
>
> But what's wrong with that?  It sounds like IL 72 is a higher-class road
> in terms of the physical road, and US 20 doesn't seem to have
> almost-motorway features.   Just because a road that is properly
> labeled primary is heavily used doesn't make it a higher class; you
> certainly wouldn't label it a motorway based on traffic count.

No, but motorways are such a special case of highway I really don't
think we should use them as a basis of comparison. You're either a
motorway, or you aren't.

>> traffic flow cares more about where the road goes, not what it looks
>> like.
>
> Sure, and routers can use that.
>
>
> Probably we need to completely decouple
>
>  nominal importance in the hierarchy of road types
>  physical characteristics
>  importance to the people who use it

Haven't we already? Physical characteristics have tags (surface,
lanes, maxspeed). It's the hierarchy that seems to be the sticking
point, and that's exactly what I thought "classification" was.

--
Kristian Zoerhoff
kristian.zoerh...@gmail.com

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

Re: [Talk-us] US highway classification

[Greg Troxel]

Kristian Zoerhoff  writes:

> On Mon, May 30, 2011 at 10:41 PM, Toby Murray  wrote:
>> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills  wrote:
>>> On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:
>>>
 I'm thinking the differences between motorways and trunks are minor.
 Trunks may have intersections, motorways don't.
>>>
>>> That's the simple way to state my opinion. It also seemed to be the thrust
>>> of most of the discussion on the talk page of the wiki page referenced
>>> previously as closest to consensus (the page itself just references the
>>> existence of the two camps and leaves it at that).
>>>
>>> In short, my position is simply that an end user expects a trunk road to be
>>> identifiably different than primary or secondary. That's how it's done on
>>> other maps, so I don't see why that's such a bad thing here.
>>
>> I agree with this as well. And I too thought this was a pretty widely
>> accepted convention.
>
> That's one accepted convention, to be sure, but it sometimes ignores
> the realities of where traffic goes.
>
> To give an example: 
>
> IL 72 (the secondary at the top of the map) is a 4- to 6-lane at-grade
> expressway; wide median, lights only every mile or so, speed limit up
> to 55 mph. It carries a fair amount of traffic, but because it
> parallels I 90 (a toll road here), it really only peaks at rush hour,
> when the toll road is near capacity..
>
> US 20 (the trunk at the map bottom), is a 4-lane, non-divided road,
> but it carries far more traffic than 72, as it connects the two
> motorways at the map ends (the Elgin-O'Hare Expressway, and the Elgin
> Bypass, which were never connected). It's not particularly
> distinguishable from a lesser 4-lane road, aside from the absurd
> amount of traffic it carries. If we stuck purely to the above
> convention, 72 would be trunk, and 20 would be primary (at best).  But

But what's wrong with that?  It sounds like IL 72 is a higher-class road
in terms of the physical road, and US 20 doesn't seem to have
almost-motorway features.   Just because a road that is properly
labeled primary is heavily used doesn't make it a higher class; you
certainly wouldn't label it a motorway based on traffic count.

> traffic flow cares more about where the road goes, not what it looks
> like.

Sure, and routers can use that.


Probably we need to completely decouple

  nominal importance in the hierarchy of road types

  physical characteristics

  importance to the people who use it


pgpWidit3jDAk.pgp
Description: PGP signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Kristian Zoerhoff]

On Mon, May 30, 2011 at 10:41 PM, Toby Murray  wrote:
> On Sun, May 29, 2011 at 4:05 PM, Nathan Mills  wrote:
>> On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:
>>
>>> I'm thinking the differences between motorways and trunks are minor.
>>> Trunks may have intersections, motorways don't.
>>
>> That's the simple way to state my opinion. It also seemed to be the thrust
>> of most of the discussion on the talk page of the wiki page referenced
>> previously as closest to consensus (the page itself just references the
>> existence of the two camps and leaves it at that).
>>
>> In short, my position is simply that an end user expects a trunk road to be
>> identifiably different than primary or secondary. That's how it's done on
>> other maps, so I don't see why that's such a bad thing here.
>
> I agree with this as well. And I too thought this was a pretty widely
> accepted convention.

That's one accepted convention, to be sure, but it sometimes ignores
the realities of where traffic goes.

To give an example: 

IL 72 (the secondary at the top of the map) is a 4- to 6-lane at-grade
expressway; wide median, lights only every mile or so, speed limit up
to 55 mph. It carries a fair amount of traffic, but because it
parallels I 90 (a toll road here), it really only peaks at rush hour,
when the toll road is near capacity..

US 20 (the trunk at the map bottom), is a 4-lane, non-divided road,
but it carries far more traffic than 72, as it connects the two
motorways at the map ends (the Elgin-O'Hare Expressway, and the Elgin
Bypass, which were never connected). It's not particularly
distinguishable from a lesser 4-lane road, aside from the absurd
amount of traffic it carries. If we stuck purely to the above
convention, 72 would be trunk, and 20 would be primary (at best).  But
traffic flow cares more about where the road goes, not what it looks
like.

-- 
Kristian Zoerhoff
kristian.zoerh...@gmail.com

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

Re: [Talk-us] US highway classification

[Toby Murray]

On Sun, May 29, 2011 at 4:05 PM, Nathan Mills  wrote:
> On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:
>
>> I'm thinking the differences between motorways and trunks are minor.
>> Trunks may have intersections, motorways don't.
>
> That's the simple way to state my opinion. It also seemed to be the thrust
> of most of the discussion on the talk page of the wiki page referenced
> previously as closest to consensus (the page itself just references the
> existence of the two camps and leaves it at that).
>
> In short, my position is simply that an end user expects a trunk road to be
> identifiably different than primary or secondary. That's how it's done on
> other maps, so I don't see why that's such a bad thing here.

I agree with this as well. And I too thought this was a pretty widely
accepted convention.

Toby

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 20:14:50 -0400, Nathan Edgars II wrote:

On 5/29/2011 8:09 PM, Nathan Mills wrote:
FSM knows the aerial imagery around here is outdated, to put it 
mildly.

Try the NAIP imagery:

http://wiki.openstreetmap.org/wiki/National_Agriculture_Imagery_Program


Not bad. The tileservers are slower than molasses and I'd rather it be 
leaf-off, but beggars can't be choosers. For some reason I thought NAIP 
was lower-res than that.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 8:09 PM, Nathan Mills wrote:

FSM knows the aerial imagery around here is outdated, to put it mildly.
Try the NAIP imagery: 
http://wiki.openstreetmap.org/wiki/National_Agriculture_Imagery_Program


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 20:00:33 -0400, Nathan Edgars II wrote:

On 5/29/2011 5:16 PM, Paul Johnson wrote:

subtle mass vandalism


This is why I ignore Paul.

Though I really wonder about this edit:
http://www.openstreetmap.org/browse/way/14751094/history


Using your standard, there's nothing to wonder about. OK-66 is fairly 
heavily traveled by people who don't want to pay the exorbitant tolls to 
get in/out of Tulsa every day. I don't specifically recall if it's an 
expressway that far north or not; it is at least up to Claremore. I 
wouldn't call it a trunk if it's not divided, but it's been the better 
part of a decade since I've driven it past Claremore, so I defer to 
people who have actually been on the road in question more recently than 
I. FSM knows the aerial imagery around here is outdated, to put it 
mildly.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 5:16 PM, Paul Johnson wrote:

subtle mass vandalism


This is why I ignore Paul.

Though I really wonder about this edit: 
http://www.openstreetmap.org/browse/way/14751094/history


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

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/29/2011 02:00 PM, Anthony wrote:
> On Sun, May 29, 2011 at 4:03 PM, Paul Johnson 
>  wrote:
>> On 05/29/2011 12:56 PM, Anthony wrote:
>>> What do you mean by "global consistency" and why is it desired?
>>
>> Having some kind of uniformity on a large scale means you wouldn't have
>> to learn how to read the map again just because you moved or traveled to
>> an arbitrary jurisdiction.
> 
> Great, so to answer your question, we've already achieved global consistency.

For motorways and POIs, sure, but for everything else, it's a real
crapshoot.  It doesn't help that we have some agitating forces
mass-replacing tags on the basis of The World According To Him, but that
appears to be a situation of subtle mass vandalism that OSM is happy to
ignore and highly resistant to actually confronting¹.

>> Also provides consistency for routers, since
>> I can pretty much guarantee there's not a router out there that
>> arbitrarily changes it's behavior based on what part of OpenStreetMap
>> it's looking at...
> 
> Then you can pretty much guarantee that there's not a router out there
> that works very well on OSM data across multiple locations.

Yeah, that's something of an issue.  We shouldn't make it harder for
data consumers to consume the data just because the human factor is
resistant to consistency.

¹ so whatever on that...it's a dead horse at this point, as much as I'd
like mass vandalism to be dealt with, it appears we're unwilling or
unable to actually handle such a situation.



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 12:09:30 -0700, Paul Johnson wrote:


I'm thinking the differences between motorways and trunks are minor.
Trunks may have intersections, motorways don't.


That's the simple way to state my opinion. It also seemed to be the 
thrust of most of the discussion on the talk page of the wiki page 
referenced previously as closest to consensus (the page itself just 
references the existence of the two camps and leaves it at that).


In short, my position is simply that an end user expects a trunk road 
to be identifiably different than primary or secondary. That's how it's 
done on other maps, so I don't see why that's such a bad thing here.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 4:03 PM, Paul Johnson  wrote:
> On 05/29/2011 12:56 PM, Anthony wrote:
>> What do you mean by "global consistency" and why is it desired?
>
> Having some kind of uniformity on a large scale means you wouldn't have
> to learn how to read the map again just because you moved or traveled to
> an arbitrary jurisdiction.

Great, so to answer your question, we've already achieved global consistency.

> Also provides consistency for routers, since
> I can pretty much guarantee there's not a router out there that
> arbitrarily changes it's behavior based on what part of OpenStreetMap
> it's looking at...

Then you can pretty much guarantee that there's not a router out there
that works very well on OSM data across multiple locations.

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

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/29/2011 12:56 PM, Anthony wrote:
> On Sun, May 29, 2011 at 3:17 PM, Paul Johnson 
>  wrote:
>> On 05/29/2011 08:37 AM, Anthony wrote:
>>> On Sun, May 29, 2011 at 10:34 AM, Richard Welty 
>>>  wrote:
>>> Now that I think about it, that's actually an excellent reason why
>>> classifications in the US should be by state.  Max speeds vary
>>> significantly by state, as well as classifications of what roads get
>>> what speeds.
>>
>> So how would that bring us closer to having global consistency, since
>> the scope of OSM is worldwide?
> 
> What do you mean by "global consistency" and why is it desired?

Having some kind of uniformity on a large scale means you wouldn't have
to learn how to read the map again just because you moved or traveled to
an arbitrary jurisdiction.  Also provides consistency for routers, since
I can pretty much guarantee there's not a router out there that
arbitrarily changes it's behavior based on what part of OpenStreetMap
it's looking at...



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 3:17 PM, Paul Johnson  wrote:
> On 05/29/2011 08:37 AM, Anthony wrote:
>> On Sun, May 29, 2011 at 10:34 AM, Richard Welty 
>>  wrote:
>> Now that I think about it, that's actually an excellent reason why
>> classifications in the US should be by state.  Max speeds vary
>> significantly by state, as well as classifications of what roads get
>> what speeds.
>
> So how would that bring us closer to having global consistency, since
> the scope of OSM is worldwide?

What do you mean by "global consistency" and why is it desired?

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

Re: [Talk-us] US highway classification

[Frederik Ramm]

Hi,

Paul Johnson wrote:

So how would that bring us closer to having global consistency, since
the scope of OSM is worldwide?


It is a common misconception that OSM should have globally consistent 
tagging standards since OSM is a world-wide project.


If someone were to really demand that, the project would come to a 
screeching halt and we'd have to set up our own version of United 
Nations bureaucracy and have them debate for 20 years before we can 
continue.


It is very well possible that the highway situation in Chile demands a 
different tagging approach than that in California, and there's no 
problem with that; we just tag it differently and that's it. In the long 
run maybe Chileans need their own rendering rules to make things look 
right for them, but that's no big deal.


You can of course do what you like in the US but I would not be 
surprised if some states had different kinks in their road network 
system that would demand a slightly different approach to tagging. And 
that wouldn't be a big deal. Diversity is one of OSM's strengths, not a 
weakness.


Bye
Frederik

--
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

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

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/29/2011 08:37 AM, Anthony wrote:
> On Sun, May 29, 2011 at 10:34 AM, Richard Welty 
>  wrote:
>> in short: a routing engine will probably use classifications where maxspeed
>> data is missing, but probably only to derive guesstimates of maxspeed
>> values.
> 
> Now that I think about it, that's actually an excellent reason why
> classifications in the US should be by state.  Max speeds vary
> significantly by state, as well as classifications of what roads get
> what speeds.

So how would that bring us closer to having global consistency, since
the scope of OSM is worldwide?



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/28/2011 06:13 PM, Nathan Mills wrote:
> On Sat, 28 May 2011 20:54:07 -0400, Nathan Edgars II wrote:
> 
>>> You described your criteria, but did not explain how trunk is more
>>> appropriate than primary for a two lane rural highway between two
>>> small-to-tiny cities. If you use trunk for that, there is no way to
>>> describe (in a way that shows up on the tiles) a road which is not a
>>> motorway but is better than the typical rural highway.
>>
>> There are many types of roads that it's not possible to describe. How
>> do you tag an unpaved classified road so the map shows that it's
>> unpaved (this is very common in the third world, but also occurs in
>> extremely rural areas of the US)? You don't.
> 
> Don't let the perfect be the enemy of the good, my friend. It's
> nonsensical to say that just because there are other roads that can't be
> adequately described (few and far between in my experience, but perhaps
> more common out west than I'm aware) we should waste trunk on US
> highways that could be adequately described with primary.

Given who you're caught in a flamewar with, is there any better response
other than, "Maybe you should try Google Mapmaker instead?"




signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Paul Johnson]

On 05/28/2011 12:19 PM, Anthony wrote:
> On Sat, May 28, 2011 at 10:40 AM, Mike N 
>  wrote:
>> On 5/28/2011 9:12 AM, Anthony wrote:
>>>
>>> Trunk has no meaning beyond "color the road the same color as other
>>> things that are tagged trunk".
>>
>>  Even color is not defined - some trunks can be toll / not toll.
>>
>>  However, trunk *could* serve as a router hint that the road is a better
>> selection than primary / secondary.
> 
> In my experience the difference between primary and trunk is generally
> very minor, to the point where I'm not sure there'd be any advantage
> at all in a router using it as a hint.

I'm thinking the differences between motorways and trunks are minor.
Trunks may have intersections, motorways don't.



signature.asc
Description: OpenPGP digital signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 12:36 PM, Richard Welty  wrote:
> On 5/29/11 11:59 AM, Anthony wrote:
>> Anyway, why argue about it?  If you have a reason to start
>> aggressively collecting data the missing maxspeed data, just do it.
>
> argue in the sense of a civil discussion of two distinctly different points
> of view is not an unreasonable thing to be doing right now, we have a
> bunch of distinctly different points of view and so far we're not entirely
> uncivil about it (although sometimes pushing it a bit.)

I'm not sure where we disagree, though.  I've got no problem with you
aggressively collecting maxspeed data.

> here is the problem with the state by state classification trick. there's
> been a standing effort do just that for a couple of years now:
>
> http://wiki.openstreetmap.org/wiki/OSM_tags_for_routing/Maxspeed#United_States_of_America
>
> as you can see, maxspeed data for exactly 3 states has been entered
> in that time frame. i'd be tempted to call this effort a dismal failure.

Do you really think Mapquest couldn't collect maxspeed information for
all 50 states in a day or two?

> there are two additional problems in addition to the obvious one that
> the tables aren't filled out and seemingly won't be:
>
> 1) getting the routing engine authors to pay attention to this data
>
> 2) the sometimes tremendous variation between highways which default
> and highways with speed limits that are at variance with these tables.
> in surveying primaries, secondaries and tertiaries in Rensselaer County
> (upstate NY) i frequently see explicit speed limits that are between
> the default 55mph and the residential presumption of 30mph.
>
> i really don't think there's any substitute for actually collecting
> maxspeed for the network of major roads. sure, try a temporary
> hack based on classifications, state-by-state if you want, but
> i suspect that effort won't be very profitable.

Well, you just said you know that MapQuest's router uses a temporary
hack based on classifications.  I would assume that hack is state by
state, since max speeds are set by the state.

> collecting maxspeed, by comparison, while very tedious and time
> consuming, goes directly to the issue at hand in routing engines.

When that effort to collect maxspeeds for all ways is completed then
we can have this discussion all over again.  Of course, by that time
we probably won't have use for roads as most people will just use
their teleporters.

> and it also allows us to focus the classification discussion on making
> "pretty" maps, which is probably where it should be focused.

I'd say "useful" maps rather than "pretty" ones, but yeah, that's
where the classification really matters.  Even beyond the maxspeed
distinction, the classification dispute only seems to be an issue when
there's no other route for miles and miles anyway.  So I don't think
the routers particularly care about this dispute.

> and the distinction we're then discussing is a simple one:
> classification as trunk based on physical characteristics vs
> perceived function in the road network. the former is much
> less subjective than the latter, but the latter does have its
> passionate supporters. but i don't see the point in discussing
> "major intercity routes" without a definition of what constitutes
> an urban area that qualifies as an endpoint or midpoint.
> that's why, for example, i'm not clear on whether NY 7/VT 9
> from Troy, NY to Bennington, VT deserves trunk rather than
> primary. at this point, i'd call it primary, but NE2 is sure it's
> trunk (incidentially, i'd probably also upgrade NY 2->MA 2
> to primary as well.)

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 10:31 AM, John F. Eldredge wrote:

Nathan Edgars II  wrote:


There are many types of roads that it's not possible to describe. How
do
you tag an unpaved classified road so the map shows that it's unpaved
(this is very common in the third world, but also occurs in extremely
rural areas of the US)? You don't.


Use the surface tag.  Using multiple tags to describe a way, instead of simply 
the highway tag, lets you describe more details.  In turn, renderers should 
look at more than just the highway tag.


I agree. It was a rhetorical question.

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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/29/11 11:59 AM, Anthony wrote:

On Sun, May 29, 2011 at 11:44 AM, Richard Welty  wrote:

On 5/29/11 11:37 AM, Anthony wrote:

On Sun, May 29, 2011 at 10:34 AM, Richard Welty
  wrote:

in short: a routing engine will probably use classifications where
maxspeed
data is missing, but probably only to derive guesstimates of maxspeed
values.

Now that I think about it, that's actually an excellent reason why
classifications in the US should be by state.  Max speeds vary
significantly by state, as well as classifications of what roads get
what speeds.


actually, i'd argue that it's an excellent reason to start aggressively
collecting the missing maxspeed data.

The two aren't mutually exclusive.

Anyway, why argue about it?  If you have a reason to start
aggressively collecting data the missing maxspeed data, just do it.

argue in the sense of a civil discussion of two distinctly different points
of view is not an unreasonable thing to be doing right now, we have a
bunch of distinctly different points of view and so far we're not entirely
uncivil about it (although sometimes pushing it a bit.)

here is the problem with the state by state classification trick. there's
been a standing effort do just that for a couple of years now:

http://wiki.openstreetmap.org/wiki/OSM_tags_for_routing/Maxspeed#United_States_of_America

as you can see, maxspeed data for exactly 3 states has been entered
in that time frame. i'd be tempted to call this effort a dismal failure.

there are two additional problems in addition to the obvious one that
the tables aren't filled out and seemingly won't be:

1) getting the routing engine authors to pay attention to this data

2) the sometimes tremendous variation between highways which default
and highways with speed limits that are at variance with these tables.
in surveying primaries, secondaries and tertiaries in Rensselaer County
(upstate NY) i frequently see explicit speed limits that are between
the default 55mph and the residential presumption of 30mph.

i really don't think there's any substitute for actually collecting
maxspeed for the network of major roads. sure, try a temporary
hack based on classifications, state-by-state if you want, but
i suspect that effort won't be very profitable.

collecting maxspeed, by comparison, while very tedious and time
consuming, goes directly to the issue at hand in routing engines.

and it also allows us to focus the classification discussion on making
"pretty" maps, which is probably where it should be focused.

and the distinction we're then discussing is a simple one:
classification as trunk based on physical characteristics vs
perceived function in the road network. the former is much
less subjective than the latter, but the latter does have its
passionate supporters. but i don't see the point in discussing
"major intercity routes" without a definition of what constitutes
an urban area that qualifies as an endpoint or midpoint.
that's why, for example, i'm not clear on whether NY 7/VT 9
from Troy, NY to Bennington, VT deserves trunk rather than
primary. at this point, i'd call it primary, but NE2 is sure it's
trunk (incidentially, i'd probably also upgrade NY 2->MA 2
to primary as well.)

richard


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 11:44 AM, Richard Welty  wrote:
> On 5/29/11 11:37 AM, Anthony wrote:
>>
>> On Sun, May 29, 2011 at 10:34 AM, Richard Welty
>>  wrote:
>>>
>>> in short: a routing engine will probably use classifications where
>>> maxspeed
>>> data is missing, but probably only to derive guesstimates of maxspeed
>>> values.
>>
>> Now that I think about it, that's actually an excellent reason why
>> classifications in the US should be by state.  Max speeds vary
>> significantly by state, as well as classifications of what roads get
>> what speeds.
>>
> actually, i'd argue that it's an excellent reason to start aggressively
> collecting the missing maxspeed data.

The two aren't mutually exclusive.

Anyway, why argue about it?  If you have a reason to start
aggressively collecting data the missing maxspeed data, just do it.

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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/29/11 11:37 AM, Anthony wrote:

On Sun, May 29, 2011 at 10:34 AM, Richard Welty  wrote:

in short: a routing engine will probably use classifications where maxspeed
data is missing, but probably only to derive guesstimates of maxspeed
values.

Now that I think about it, that's actually an excellent reason why
classifications in the US should be by state.  Max speeds vary
significantly by state, as well as classifications of what roads get
what speeds.


actually, i'd argue that it's an excellent reason to start aggressively
collecting the missing maxspeed data.

richard


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 10:34 AM, Richard Welty  wrote:
> in short: a routing engine will probably use classifications where maxspeed
> data is missing, but probably only to derive guesstimates of maxspeed
> values.

Now that I think about it, that's actually an excellent reason why
classifications in the US should be by state.  Max speeds vary
significantly by state, as well as classifications of what roads get
what speeds.

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

Re: [Talk-us] US highway classification

[John F. Eldredge]

Nathan Edgars II  wrote:

>There are many types of roads that it's not possible to describe. How
>do 
>you tag an unpaved classified road so the map shows that it's unpaved 
>(this is very common in the third world, but also occurs in extremely 
>rural areas of the US)? You don't.

Use the surface tag.  Using multiple tags to describe a way, instead of simply 
the highway tag, lets you describe more details.  In turn, renderers should 
look at more than just the highway tag.

-- 
John F. Eldredge -- j...@jfeldredge.com
"Reserve your right to think, for even to think wrongly is better than not to 
think at all." -- Hypatia of Alexandria

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 03:00:03 -0400, Nathan Edgars II wrote:


Perhaps the best way to handle it would be to render a wider line if
oneway=yes and not lanes=1 or if oneway=no/unset and lanes=4 or more.
Thus divided highways would not need a lane count to be wider, but
undivided roads would need to be tagged as having four lanes.


That seems like it would be a reasonable way to handle it. I'm no 
mapnik expert, but I can see what I can figure out with the installation 
I have locally. I don't think that tagging all divided highways that 
aren't motorways trunk would be a terribly bad idea at first blush, 
though.



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

Re: [Talk-us] US highway classification

[Chris Lawrence]

On Fri, May 27, 2011 at 12:32 AM, Nathan Mills  wrote:
> Would I be correct in stating that tagging an undivided 2 lane (one lane in
> each direction) highways would be improper, even if a state calls the
> highway a "trunk" for planning purposes?

No, you wouldn't.  Trunk is the proper classification for major
through routes that lack full access control in OSM, regardless of
their number of lanes, median (or lack thereof) etc.  For example, US
59 from Laredo to Houston varies between a behemoth of a freeway with
HOV lanes and a two-lane road with 30 mph speed limits in places, yet
it's a major through route throughout for intercity car, bus, and
truck traffic.

Moreover routes shouldn't be semi-randomly changing classification due
to speed limits, lane width, etc.  So just passing through a 20 mph
zone doesn't demote a road to residential status, and neither does a
75 mph Texas Farm-to-Market route become a trunk just because its
speed limit is higher than the legal speed limit in most other states.

In any event, once upon a time there were clear and consistent
definitions in the Wiki for the USA (I believe the ones currently in
"United States roads tagging" are the closest to those)... and then
they got duplicated in different places in the wiki willy-nilly and
there are now a proliferation of silly ones that are often
contradictory.

That said if there are isolated bits of routes that aren't major
through routes that have gotten upgraded to trunk, by all means demote
them back to their appropriate level of importance.


Chris

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 2:30 AM, Nathan Mills wrote:

I think that trunk is more useful if it's prescriptive, more along the
lines of a motorway than primary and below. If we aren't going to do
that, we need to come up with another value for highway and get it
rendered by default. It's something that map users expect, and we should
therefore deliver.


How many maps actually show these with a separate symbol, though? 
Looking through my 1990s road atlases, AAA and National Geographic 
(MapQuest) simply modify the symbol, keeping the color of the two-lane 
segments, while Rand McNally and Gousha (RIP) use one symbol for all 
divided highways, no matter how important (in Orlando, this includes 
Colonial, 17-92, OBT, and Michigan). (All four count center turn lanes 
as medians - see South OBT.) Unless you're proposing to mark all divided 
highways, rural and urban, as trunk, map users accustomed to road 
atlases won't expect your criteria.


Perhaps the best way to handle it would be to render a wider line if 
oneway=yes and not lanes=1 or if oneway=no/unset and lanes=4 or more. 
Thus divided highways would not need a lane count to be wider, but 
undivided roads would need to be tagged as having four lanes.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 02:18:09 -0400, Nathan Edgars II wrote:

On 5/29/2011 1:50 AM, Nathan Mills wrote:


It's actually faster to take 441 to Yeehaw and get on the turnpike 
there
when traveling from eastern and southeastern Orlando to points south 
of

Port St. Lucie.

Even with the four-laning of 192?


Yes. That more people don't do it is just them being silly. Good for 
me, I suppose.



Speaking of misclassification around Orlando, why on did you make
Alafaya Trail south of Curry Ford primary?

To distinguish it from the adjacent secondaries, which are similarly
more major than the tertiaries. It's a balancing act, not an exact
science.


Indeed, but by any standard Lake Underhill and Curry Ford are both more 
heavily trafficked and more important than that segment of Alafaya. (for 
now)



We're obviously getting nowhere here. You think trunk should be used
for certain physical characteristics, and other people think it 
should

be used for a slightly different set. I think a more systematic
approach makes sense, classifying the most major routes in the system
as trunk. Again, even under that view, there will be disagreement 
over
where the line is drawn. But you seem to be rejecting that it's even 
a

valid option, like if someone were to insist that primaries must have
at least four lanes, or that tertiaries must have a centerline.


I think that trunk is more useful if it's prescriptive, more along the 
lines of a motorway than primary and below. If we aren't going to do 
that, we need to come up with another value for highway and get it 
rendered by default. It's something that map users expect, and we should 
therefore deliver. Even with your view, physical characteristics are 
part of what makes a particular route important within the network. TBH, 
I think you overuse the trunk tag regardless of physical 
characteristics.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 1:50 AM, Nathan Mills wrote:

On Sun, 29 May 2011 01:00:25 -0400, Nathan Edgars II wrote:

On 5/29/2011 12:37 AM, Nathan Mills wrote:

US-441 between St. Cloud and Yeehaw Junction could easily be trunk by
NE2's definition

Nope, since any through traffic will be on the Turnpike. US 441
serves mainly only local and toll-avoiding traffic, and the latter is
better-off cutting east to I-95 via US 192.


It's actually faster to take 441 to Yeehaw and get on the turnpike there
when traveling from eastern and southeastern Orlando to points south of
Port St. Lucie.
Even with the four-laning of 192? That's still a relatively small amount 
of traffic - people from a certain part of Orlando (eastern but not too 
far east) who want to go south. The principal route from Orlando to 
Miami is obviously the Turnpike, even if a there's a better route from a 
few parts of the Orlando area.



Speaking of misclassification around Orlando, why on did you make
Alafaya Trail south of Curry Ford primary?
To distinguish it from the adjacent secondaries, which are similarly 
more major than the tertiaries. It's a balancing act, not an exact science.



We're obviously getting nowhere here. You think trunk should be used for 
certain physical characteristics, and other people think it should be 
used for a slightly different set. I think a more systematic approach 
makes sense, classifying the most major routes in the system as trunk. 
Again, even under that view, there will be disagreement over where the 
line is drawn. But you seem to be rejecting that it's even a valid 
option, like if someone were to insist that primaries must have at least 
four lanes, or that tertiaries must have a centerline.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 01:00:25 -0400, Nathan Edgars II wrote:

On 5/29/2011 12:37 AM, Nathan Mills wrote:
US-441 between St. Cloud and Yeehaw Junction could easily be trunk 
by

NE2's definition

Nope, since any through traffic will be on the Turnpike. US 441
serves mainly only local and toll-avoiding traffic, and the latter is
better-off cutting east to I-95 via US 192.


It's actually faster to take 441 to Yeehaw and get on the turnpike 
there when traveling from eastern and southeastern Orlando to points 
south of Port St. Lucie. Once again, this is a drive I take 
semi-regularly. Similarly, from northeast Orlando out to Cocoa, 520 is 
faster than slogging all the way down to 528 on 417, although almost 
nobody goes that way. (or at least they didn't when I was driving it 
somewhat regularly..it's been some years at this point)


On all the roads I've mentioned until this post (exclusive of US-71), 
most of the traffic seemed to be local and there was very little 
traffic, local or through. So that's not really something you can point 
to to say "aha, this should be trunk because it's an important through 
route" on the particular roads I'm discussing. I don't claim to know 
about roads I haven't personally driven since I haven't actually been on 
them. If you can divine from the Internet that a given road has lots of 
through traffic, that's fantastic for you.


Speaking of misclassification around Orlando, why on did you make 
Alafaya Trail south of Curry Ford primary? Even with the extension 
(Innovation way) to 528, it's still mostly local traffic. Curry Ford and 
Lake Underhill are much more important (and busy) roads in that part of 
Orlando and carry more through traffic. That may change if they ever get 
the 2 lane segment 4 laned, but that hasn't happened yet, unless it was 
the fastest construction project evar. (I have friends that live in 
Stoneybrook whom I visit at least once a year; last time was admittedly 
last Christmas. I can only put so many miles on the car every year :P)


On Sun, 29 May 2011 01:16:17 -0400, Nathan Edgars II wrote:


Try expressway=yes or access_control=partial. But make sure it's an
expressway, not just a four-lane with unlimited driveway access. You
may have to look through DOT records to find out which one it is.


I don't think most road map users care whether it's a literal 
expressway in the sense of limited access or not, so long as it is 
practically-speaking indistinguishable from an actual limited access 
road. This is why I'm agitating for a tag that captures that use. More 
technical tagging is, of course, quite welcome, but we need a way to 
represent something as basic as this on the map. I think the easiest way 
is to use the trunk tag because I don't see the way you're using it as 
particularly useful since the same information could be conveyed with 
primary, even if a nearby road has to be demoted to secondary, as most 
state highways already are or should be, IMO.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 1:08 AM, Nathan Mills  wrote:
> On Sun, 29 May 2011 00:57:30 -0400, Anthony wrote:
>
>>> That's quite the misrepresentation of what I'm saying.
>>
>> It was an exact quote.
>
> You may have heard of the concept of the "pull quote." It describes using
> partial quotations to misrepresent someone else's position.

Yes, I have.

>> And my point is 1) that you aren't going to convince people to do
>> that; and 2) that if you could convince people to tag the number of
>> lanes, you'd be better off having them use a tag which says the number
>> of lanes.
>
> I find it difficult to believe that you object so strenuously to making it
> simple to tag one of the main things an end user of a road map desires to
> know when looking at said map.

Good, because I don't.

> Is it a practical "you can't get people to
> agree to that" objection, or a "I don't think it should be done that way"
> objection?

It's both.  I don't think you are going to convince people to base the
trunk tag on the number of lanes, and I don't think we should use the
term trunk to tag the fact that a road has four lanes.

> Once again, there is, to most non-mapgeeks a class of road which is less
> than a motorway, but better than all other classes of road. In my part of
> the country, most people call it an expressway. This should be easy to tag,
> so that the map is most useful to end users (and simple to edit for casual
> editors, who you're almost certainly not going to convince to tag width and
> lane count on every edit).

I agree.

> Trunk seems to fit that bill, and is used that way already in many areas.

Here I disagree.  The problem with using trunk to represent the class
of road, and primary to represent the most important non-motorway, is
that you can't tag something trunk *and* primary.

> What advantage does trunk have over primary in any of the mentioned
> examples?

It is rendered at z5.

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 1:08 AM, Nathan Mills wrote:

Once again, there is, to most non-mapgeeks a class of road which is less
than a motorway, but better than all other classes of road. In my part
of the country, most people call it an expressway.


Try expressway=yes or access_control=partial. But make sure it's an 
expressway, not just a four-lane with unlimited driveway access. You may 
have to look through DOT records to find out which one it is.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 01:04:24 -0400, Anthony wrote:
On Sun, May 29, 2011 at 12:59 AM, Nathan Mills  
wrote:

On Sun, 29 May 2011 00:13:33 -0400, Anthony wrote:


convincing people that their
area of the country isn't allowed to have any trunks.


Also, why is this any worse than not having a motorway?


Why is what worse than not having a motorway?


Not having any trunks. Why is that worse than not having any motorways?


I don't think the
folks in Newton County Arkansas care a whit whether the main road 
through
their very rural county is trunk or primary or secondary or even 
tertiary.


They do if they're looking at a map, and none of the roads are drawn,
because tertiaries aren't drawn at that zoom level.


I don't think it's a problem if rural areas have a low density of roads 
drawn at low zoom. That's how most every map made is drawn. (Yes, 
tertiary may have been taking it too far, but for reasons of rendering, 
not the importance of the road)


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 00:57:30 -0400, Anthony wrote:


That's quite the misrepresentation of what I'm saying.


It was an exact quote.


You may have heard of the concept of the "pull quote." It describes 
using partial quotations to misrepresent someone else's position.



Again, my point is
that trunk is much more useful (especially to people using rendered 
mapnik
tiles) if it is mainly restricted to four lane divided sorts of 
roads here

in the US.


And my point is 1) that you aren't going to convince people to do
that; and 2) that if you could convince people to tag the number of
lanes, you'd be better off having them use a tag which says the 
number

of lanes.


I find it difficult to believe that you object so strenuously to making 
it simple to tag one of the main things an end user of a road map 
desires to know when looking at said map. Is it a practical "you can't 
get people to agree to that" objection, or a "I don't think it should be 
done that way" objection?


Once again, there is, to most non-mapgeeks a class of road which is 
less than a motorway, but better than all other classes of road. In my 
part of the country, most people call it an expressway. This should be 
easy to tag, so that the map is most useful to end users (and simple to 
edit for casual editors, who you're almost certainly not going to 
convince to tag width and lane count on every edit). Trunk seems to fit 
that bill, and is used that way already in many areas. It was used that 
way in a lot more areas until one specific editor decided he wanted to 
edit roads in places he's never even been to use that designation. I 
can't think of a downside to using it that way.


What advantage does trunk have over primary in any of the mentioned 
examples?


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 12:59 AM, Nathan Mills  wrote:
> On Sun, 29 May 2011 00:13:33 -0400, Anthony wrote:
>
>> convincing people that their
>> area of the country isn't allowed to have any trunks.
>
> Also, why is this any worse than not having a motorway?

Why is what worse than not having a motorway?

> I don't think the
> folks in Newton County Arkansas care a whit whether the main road through
> their very rural county is trunk or primary or secondary or even tertiary.

They do if they're looking at a map, and none of the roads are drawn,
because tertiaries aren't drawn at that zoom level.

On Sun, May 29, 2011 at 12:37 AM, Nathan Mills  wrote:
> US-441 between St. Cloud and Yeehaw Junction could easily be trunk by NE2's
> definition (and apparently yours, although you haven't really indicated how
> exactly it should be used)

Having now looked at it quickly on other maps, yes, I think it could.

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/29/2011 12:37 AM, Nathan Mills wrote:

US-441 between St. Cloud and Yeehaw Junction could easily be trunk by
NE2's definition
Nope, since any through traffic will be on the Turnpike. US 441 serves 
mainly only local and toll-avoiding traffic, and the latter is 
better-off cutting east to I-95 via US 192.


Now US 441 south of Yeehaw might be, given that it provides access to 
the entire east side of Lake Okeechobee from central Florida (and so 
FDOT has put it on their Statewide Intermodal System). But in Florida I 
tried to stick with rural four-lanes and their continuations.


A decent example of parallel trunk and primary can be seen east of 
Gainesville. Both SR 20 and SR 26-100 connect Gainesville to Palatka, 
and are about the same length. But SR 200 is the better route, having 
been partially four-laned. (Yes, SR 26 could be secondary, but it's more 
major than the average secondary.)


> primary rather than trunk (as many of them were for years before NE2

went off and changed them).
Since most of those were primary from the TIGER import, that's not a 
very strong argument for keeping them primary.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 00:13:33 -0400, Anthony wrote:


convincing people that their
area of the country isn't allowed to have any trunks.


Also, why is this any worse than not having a motorway? I don't think 
the folks in Newton County Arkansas care a whit whether the main road 
through their very rural county is trunk or primary or secondary or even 
tertiary. The roads in that county just aren't major roads. That's not 
disparaging the people, it's just a statement of what is.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sun, May 29, 2011 at 12:37 AM, Nathan Mills  wrote:
> On Sun, 29 May 2011 00:13:33 -0400, Anthony wrote:
>> If you want to get people to tag "more than two lanes" and "a
>> barely-existent shoulder", I think you'd have much more success
>> creating tags for those features than convincing people that their
>> area of the country isn't allowed to have any trunks.
>
> That's quite the misrepresentation of what I'm saying.

It was an exact quote.

> Again, my point is
> that trunk is much more useful (especially to people using rendered mapnik
> tiles) if it is mainly restricted to four lane divided sorts of roads here
> in the US.

And my point is 1) that you aren't going to convince people to do
that; and 2) that if you could convince people to tag the number of
lanes, you'd be better off having them use a tag which says the number
of lanes.

On Sun, May 29, 2011 at 12:37 AM, Nathan Mills  wrote:
> US-441 between St. Cloud and Yeehaw Junction could easily be trunk by NE2's
> definition (and apparently yours, although you haven't really indicated how
> exactly it should be used)

I indicated how trunk should be used in my first post to this thread.
When you've got a non-motorway, and you want mapnik to color the road
green, and show it at z5, you use trunk.

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sun, 29 May 2011 00:13:33 -0400, Anthony wrote:


If you want to get people to tag "more than two lanes" and "a
barely-existent shoulder", I think you'd have much more success
creating tags for those features than convincing people that their
area of the country isn't allowed to have any trunks.


That's quite the misrepresentation of what I'm saying. Again, my point 
is that trunk is much more useful (especially to people using rendered 
mapnik tiles) if it is mainly restricted to four lane divided sorts of 
roads here in the US. You can bring up all the corner cases you want, 
but that doesn't change the fact that using trunk to describe roads that 
can adequately handled with primary (for all purposes, no less!) means 
that it's simply not possible to represent that with any tagging simple 
enough to be reliably used everywhere.


US-441 between St. Cloud and Yeehaw Junction could easily be trunk by 
NE2's definition (and apparently yours, although you haven't really 
indicated how exactly it should be used), but is also adequately 
described by primary, and handily enough renders rather well.


Yeah, we shouldn't tag specifically for the renderer, but we shouldn't 
waste tags either. Nor should we avoid being mindful of how things are 
rendered.


Perhaps if you explain it very slowly, you can help me understand why 
primary isn't emphatic enough in the cases that have been mentioned. Or 
how it is that a routing engine would be confused by them being tagged 
as primary rather than trunk (as many of them were for years before NE2 
went off and changed them). As I said, I see those changes as making the 
map less useful to someone reading it, not more useful.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 11:29 PM, Nathan Mills  wrote:
> Obviously, my preference would be that trunk only be used for roads with
> more than two lanes and a barely-existent shoulder, but barring that I would
> like it to mean 'very important road that is not a motorway.'
>
> And yes, as it presently stands, a routing engine would probably be better
> off using the two foot shorter primary than the two foot longer trunk, given
> that we have a bunch of roads tagged as trunk that are no more suitable for
> long distance travel than most roads tagged as primary.

Within the same geographic area?  I mean sure, there's a trunk in
Grove Hill, Alabama which is worse than a primary in Tampa, Florida.
But so long as there aren't any trunks in Grove Hill, Alabama which
are worse than primaries in Grove Hill, Alabama, I think the routers
and renderers should do okay.

If you want to get people to tag "more than two lanes" and "a
barely-existent shoulder", I think you'd have much more success
creating tags for those features than convincing people that their
area of the country isn't allowed to have any trunks.

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 10:52 PM, Nathan Mills wrote:

Only if trunk has a meaning that implies that a road tagged trunk is
somehow better than a road tagged primary, which it apparently does not,
at least in some people's minds. If you're going to waste trunk on curvy
two lane roads, a router may as well use distance or maxspeed as a
better metric. As it stands, some of us are using trunk as more of a 'I
know it when I see it' thing than something useful for routing purposes
like motorway.


Your same argument applies to the difference between primary and 
secondary, and that between secondary and tertiary.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 23:00:11 -0400, Anthony wrote:


Instead of giving me hypothetical if..then answers, can you give me a
straightforward answer?


You're trying to get an exact answer to something that isn't an exact 
science, so no. I'm allowing for the fact that there may be a situation 
in which trunk should be applied to a two lane road because not doing so 
would misrepresent the area to viewers of the map and routing engines 
more than not using it would. I don't have personal experience with 
every road in the US, after all. In the cases I've personally seen, I 
think the roads could have been adequately described with primary.


Obviously, my preference would be that trunk only be used for roads 
with more than two lanes and a barely-existent shoulder, but barring 
that I would like it to mean 'very important road that is not a 
motorway.'


And yes, as it presently stands, a routing engine would probably be 
better off using the two foot shorter primary than the two foot longer 
trunk, given that we have a bunch of roads tagged as trunk that are no 
more suitable for long distance travel than most roads tagged as 
primary.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 10:49 PM, Nathan Mills  wrote:
> On Sat, 28 May 2011 22:39:51 -0400, Anthony wrote:
>>
>> On Sat, May 28, 2011 at 9:47 PM, Nathan Mills  wrote:
>>>
>>> Primary means (at least according to most of the wiki pages)
>>> the primary non-motorway route between two cities.
>>
>> Any wiki pages that say that are clearly wrong.  Trunk is the primary
>> non-motorway route between two cities.  Yes, it's dumb terminology,
>> but it's too late to fix that.
>
> If it were clearly wrong, we wouldn't be having this discussion.

It is, and we are.

>>> Another example is US-71 between Fort Smith and Texarkana. It is in fact
>>> the
>>> fastest route between Fort Smith and Texarkana, but it is terribly slow
>>> going. The fact that it is the fastest route between those two regionally
>>> important cities is adequately described by primary. Why, then, are we
>>> wasting trunk on something like that?
>>
>> Do you agree that a road which would be a trunk in one area of the
>> world might not be a trunk in another area?  Or do I first have to
>> convince you of that?
>
> I said as much previously. Obviously, I'm only considering the US, given
> that this is talk-us.

Right, which leads to my next question.  Do you agree that a road
which would be a trunk in one area of the US might not be a trunk in
another area?

Grove Hill, Alabama, and New York City, New York don't exactly have
very much in common.

> I guess I'm just failing to see what use trunk is if it's essentially 
> interchangeable with primary.

It's not essentially interchangeable with primary.

On Sat, May 28, 2011 at 10:52 PM, Nathan Mills  wrote:
>> You agree that if a router has two possible roads to take between two
>> cities, and one is a trunk, and one is a primary, and all other things
>> are equal, that the router should choose the trunk, right?  Doesn't
>> that make trunk, by definition, the primary non-motorway route between
>> two cities?
>
> Only if trunk has a meaning that implies that a road tagged trunk is somehow
> better than a road tagged primary, which it apparently does not, at least in
> some people's minds.

So no, routers shouldn't choose trunk over primary, all other things
being equal?

> If you're going to waste trunk on curvy two lane roads,
> a router may as well use distance or maxspeed as a better metric.

We're assuming maxspeed isn't present.

So, if the primary road is 2 inches shorter than the trunk, the router
should take the primary?

And the reasoning for this is, to spite the people who "wasted" trunk
on curvy two lane roads (when there was nothing better in a 50 mile
radius)?

Instead of giving me hypothetical if..then answers, can you give me a
straightforward answer?

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

Re: [Talk-us] US highway classification

[Nathan Mills]

You agree that if a router has two possible roads to take between two
cities, and one is a trunk, and one is a primary, and all other 
things

are equal, that the router should choose the trunk, right?  Doesn't
that make trunk, by definition, the primary non-motorway route 
between

two cities?


Only if trunk has a meaning that implies that a road tagged trunk is 
somehow better than a road tagged primary, which it apparently does not, 
at least in some people's minds. If you're going to waste trunk on curvy 
two lane roads, a router may as well use distance or maxspeed as a 
better metric. As it stands, some of us are using trunk as more of a 'I 
know it when I see it' thing than something useful for routing purposes 
like motorway.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 22:39:51 -0400, Anthony wrote:
On Sat, May 28, 2011 at 9:47 PM, Nathan Mills  
wrote:

Primary means (at least according to most of the wiki pages)
the primary non-motorway route between two cities.


Any wiki pages that say that are clearly wrong.  Trunk is the primary
non-motorway route between two cities.  Yes, it's dumb terminology,
but it's too late to fix that.


If it were clearly wrong, we wouldn't be having this discussion.

Another example is US-71 between Fort Smith and Texarkana. It is in 
fact the
fastest route between Fort Smith and Texarkana, but it is terribly 
slow
going. The fact that it is the fastest route between those two 
regionally
important cities is adequately described by primary. Why, then, are 
we

wasting trunk on something like that?


Do you agree that a road which would be a trunk in one area of the
world might not be a trunk in another area?  Or do I first have to
convince you of that?


I said as much previously. Obviously, I'm only considering the US, 
given that this is talk-us. I guess I'm just failing to see what use 
trunk is if it's essentially interchangeable with primary.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 10:39 PM, Anthony  wrote:
> On Sat, May 28, 2011 at 9:47 PM, Nathan Mills  wrote:
>> Primary means (at least according to most of the wiki pages)
>> the primary non-motorway route between two cities.
>
> Any wiki pages that say that are clearly wrong.  Trunk is the primary
> non-motorway route between two cities.  Yes, it's dumb terminology,
> but it's too late to fix that.

You agree that if a router has two possible roads to take between two
cities, and one is a trunk, and one is a primary, and all other things
are equal, that the router should choose the trunk, right?  Doesn't
that make trunk, by definition, the primary non-motorway route between
two cities?

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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 9:47 PM, Nathan Mills  wrote:
> Primary means (at least according to most of the wiki pages)
> the primary non-motorway route between two cities.

Any wiki pages that say that are clearly wrong.  Trunk is the primary
non-motorway route between two cities.  Yes, it's dumb terminology,
but it's too late to fix that.

> Another example is US-71 between Fort Smith and Texarkana. It is in fact the
> fastest route between Fort Smith and Texarkana, but it is terribly slow
> going. The fact that it is the fastest route between those two regionally
> important cities is adequately described by primary. Why, then, are we
> wasting trunk on something like that?

Do you agree that a road which would be a trunk in one area of the
world might not be a trunk in another area?  Or do I first have to
convince you of that?

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 21:51:31 -0400, Nathan Edgars II wrote:


It's been rebuilt as a good-quality four-lane in Mississippi, eastern
Alabama, and Georgia. Alabama has been a little slower at four-laning
than its neighbors, but US 84 in western Alabama is still a direct
route connecting the four-laned portions.


More than a little slower. There's perhaps 20 miles of four lane in 
Alabama west of Enterprise. Maybe 30 if they've completed the far 
western section they were working on last time I was through there. As I 
mentioned in another post, it's faster to detour to I-10. And once 
again, the 'direct route' issue is adequately communicated with the 
primary tag.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 9:47 PM, Nathan Mills wrote:

Another example is US-71 between Fort Smith and Texarkana. It is in fact
the fastest route between Fort Smith and Texarkana, but it is terribly
slow going. The fact that it is the fastest route between those two
regionally important cities is adequately described by primary. Why,
then, are we wasting trunk on something like that?
I could ask you the same question: why are we wasting primary on what 
can be adequately described by trunk?


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 9:13 PM, Nathan Mills wrote:

So you continue to assert that trunk is most useful if it essentially a
duplicate of primary?

Maybe a duplicate of your version of primary, but not mine.


Take, as an example, US 84 in western Alabama. Why on earth did you
change it to trunk when it's a terribly substandard road that isn't even
a major route between cities of any real size?


It's been rebuilt as a good-quality four-lane in Mississippi, eastern 
Alabama, and Georgia. Alabama has been a little slower at four-laning 
than its neighbors, but US 84 in western Alabama is still a direct route 
connecting the four-laned portions.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 21:30:50 -0400, Anthony wrote:


Say, Dothan, Alabama to Hattiesburg, Mississippi, avoid motorways.
What should the router take?


In that particular case, it should in fact take US-84. (US-231 to I-10 
to US-98 would in fact be faster; I know this having taken both routes, 
but you said avoid motorways) But US-84 ought to be tagged primary for 
the reasons I previously stated. Trunk is useless if it describes the 
same thing that primary does. Primary means (at least according to most 
of the wiki pages) the primary non-motorway route between two cities. It 
does not imply any particular quality of road. That seems perfectly 
applicable to US-84 in most of central and western Alabama.


Another example is US-71 between Fort Smith and Texarkana. It is in 
fact the fastest route between Fort Smith and Texarkana, but it is 
terribly slow going. The fact that it is the fastest route between those 
two regionally important cities is adequately described by primary. Why, 
then, are we wasting trunk on something like that? Trunk seems more 
appropriate for OK-51 between I-35 and Stillwater or the non-turnpike 
(and non-city) sections of US-412 between Tulsa, OK and Springdale, AR 
where it has a meaning above and beyond what primary means.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 9:27 PM, Anthony  wrote:
> On Sat, May 28, 2011 at 9:13 PM, Nathan Mills  wrote:
>> Take, as an example, US 84 in western Alabama.
>
> FWIW, Google has it as the top level non-motorway.  As far as I can
> tell there's no other more important east-west road within 50 miles.
>
> What road would you use traveling east/west along the same route?

Say, Dothan, Alabama to Hattiesburg, Mississippi, avoid motorways.
What should the router take?

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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 9:13 PM, Nathan Mills  wrote:
> Take, as an example, US 84 in western Alabama.

FWIW, Google has it as the top level non-motorway.  As far as I can
tell there's no other more important east-west road within 50 miles.

What road would you use traveling east/west along the same route?

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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 3:39 PM, Nathan Mills  wrote:
> On Sat, 28 May 2011 15:19:03 -0400, Anthony wrote:
>
>> In my experience the difference between primary and trunk is generally
>> very minor, to the point where I'm not sure there'd be any advantage
>> at all in a router using it as a hint.
>>
>> But maybe that's just because the places where I use OSM are mapped wrong.
>
> Using NE2's criteria, trunk is not really any different from a routing
> standpoint than primary. Using mine, trunk is something that would be
> primary if it weren't a physically better sort of road, so should be
> preferred over a nearby primary if the distances are similar. I'm harping on
> physical characteristics for this one particular tag because I think without
> that as a differentiator, it's basically useless fluff we shouldn't be using
> at all because primary already covers that usage.

For routing purposes, I think trunk is destined to be useless fluff
(even if your ideas about how to tag it so it isn't were perfect,
you'd still have the problem of convincing the mappers to use it).
For rendering purposes, I think there's room for a road, which isn't a
motorway, which shows up at least one zoom higher (less zoomed) than
primary.  I suppose that's what trunk is.

In theory maybe the rendering software could somehow figure that out
itself, but I'm not sure how feasible that is right this moment.

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 20:54:07 -0400, Nathan Edgars II wrote:


You described your criteria, but did not explain how trunk is more
appropriate than primary for a two lane rural highway between two
small-to-tiny cities. If you use trunk for that, there is no way to
describe (in a way that shows up on the tiles) a road which is not a
motorway but is better than the typical rural highway.


There are many types of roads that it's not possible to describe. How
do you tag an unpaved classified road so the map shows that it's
unpaved (this is very common in the third world, but also occurs in
extremely rural areas of the US)? You don't.


Don't let the perfect be the enemy of the good, my friend. It's 
nonsensical to say that just because there are other roads that can't be 
adequately described (few and far between in my experience, but perhaps 
more common out west than I'm aware) we should waste trunk on US 
highways that could be adequately described with primary.





I also upgrade major state-numbered highways from secondary to
primary. This leaves more breathing room in secondary and tertiary 
for

the lesser roads.


As makes sense if the highway is the most direct non-Interstate,
non-trunk route between two regionally important cities. Why would 
trunk

be used for the same thing? That's what I've been trying (apparently
rather poorly) to get at.


I understand your assumption - that trunk is only to be used for
surface expressways. I simply disagree.


So you continue to assert that trunk is most useful if it essentially a 
duplicate of primary?


Take, as an example, US 84 in western Alabama. Why on earth did you 
change it to trunk when it's a terribly substandard road that isn't even 
a major route between cities of any real size? The westernmost couple of 
miles in Alabama were being upgraded last I was there, but the rest of 
it from Elba on west is pretty bad, excluding the bypasses, of course. 
There is no reasonable standard that could have it classified the same 
as 231 between Dothan and Montgomery or even 280 between Birmingham and 
Columbus, GA, now that it's almost 100% upgraded and almost all of the 
towns are bypassed. It makes zero sense to a user of the map. (of which 
I count myself..I've been eating my own dog food for a few years now, 
which is why I care in more than an abstract way)


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 8:37 PM, Nathan Mills wrote:

On Sat, 28 May 2011 17:25:17 -0400, Nathan Edgars II wrote:


No, trunk is to primary as primary is to secondary.


Except that it's not.


It is in my criteria, which you're misrepresenting.


You described your criteria, but did not explain how trunk is more
appropriate than primary for a two lane rural highway between two
small-to-tiny cities. If you use trunk for that, there is no way to
describe (in a way that shows up on the tiles) a road which is not a
motorway but is better than the typical rural highway.


There are many types of roads that it's not possible to describe. How do 
you tag an unpaved classified road so the map shows that it's unpaved 
(this is very common in the third world, but also occurs in extremely 
rural areas of the US)? You don't.



I also upgrade major state-numbered highways from secondary to
primary. This leaves more breathing room in secondary and tertiary for
the lesser roads.


As makes sense if the highway is the most direct non-Interstate,
non-trunk route between two regionally important cities. Why would trunk
be used for the same thing? That's what I've been trying (apparently
rather poorly) to get at.


I understand your assumption - that trunk is only to be used for surface 
expressways. I simply disagree.


Whose route network a given highway is a part of seems to me to be a
poor differentiator.
Agreed. It's simply one of many data points to be taken into account, 
just like (in my opinion) physical characteristics.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 17:25:17 -0400, Nathan Edgars II wrote:


No, trunk is to primary as primary is to secondary.


Except that it's not.


It is in my criteria, which you're misrepresenting.


You described your criteria, but did not explain how trunk is more 
appropriate than primary for a two lane rural highway between two 
small-to-tiny cities. If you use trunk for that, there is no way to 
describe (in a way that shows up on the tiles) a road which is not a 
motorway but is better than the typical rural highway.



I also upgrade major state-numbered highways from secondary to
primary. This leaves more breathing room in secondary and tertiary 
for

the lesser roads.


As makes sense if the highway is the most direct non-Interstate, 
non-trunk route between two regionally important cities. Why would trunk 
be used for the same thing? That's what I've been trying (apparently 
rather poorly) to get at.


Whose route network a given highway is a part of seems to me to be a 
poor differentiator. A city-maintained motorway is the same as a 
state-maintained motorway, IMO. As I said before, some element of 
judgement is necessary in deciding whether a four lane divided highway 
really deserves trunk or whether a two lane highway really deserves 
primary. If it's not regionally important and/or is a very short 
segment, my answer would be no.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 5:04 PM, Nathan Mills wrote:

On Sat, 28 May 2011 16:21:24 -0400, Nathan Edgars II wrote:


Using NE2's criteria, trunk is not really any different from a routing
standpoint than primary.


No, trunk is to primary as primary is to secondary.


Except that it's not.


It is in my criteria, which you're misrepresenting.


You have upgraded roads from primary to trunk
where there are no roads tagged primary in the vicinity and that are not
physically any better than the usual rural primary. You're wasting a
classification by making a distinction where there is no real
difference.
I also upgrade major state-numbered highways from secondary to primary. 
This leaves more breathing room in secondary and tertiary for the lesser 
roads.


> Trunk would be better used for the next class of highway

below motorway, not the arbitrary standard you seem to be using.


I'm not sure why you persist in calling my standard arbitrary when your 
standard for primary and lower are probably just as arbitrary.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 16:21:24 -0400, Nathan Edgars II wrote:

Using NE2's criteria, trunk is not really any different from a 
routing

standpoint than primary.


No, trunk is to primary as primary is to secondary.


Except that it's not. You have upgraded roads from primary to trunk 
where there are no roads tagged primary in the vicinity and that are not 
physically any better than the usual rural primary. You're wasting a 
classification by making a distinction where there is no real 
difference. Trunk would be better used for the next class of highway 
below motorway, not the arbitrary standard you seem to be using.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/28/2011 3:39 PM, Nathan Mills wrote:

On Sat, 28 May 2011 15:19:03 -0400, Anthony wrote:


In my experience the difference between primary and trunk is generally
very minor, to the point where I'm not sure there'd be any advantage
at all in a router using it as a hint.

But maybe that's just because the places where I use OSM are mapped
wrong.


Using NE2's criteria, trunk is not really any different from a routing
standpoint than primary.


No, trunk is to primary as primary is to secondary.

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 15:19:03 -0400, Anthony wrote:

In my experience the difference between primary and trunk is 
generally

very minor, to the point where I'm not sure there'd be any advantage
at all in a router using it as a hint.

But maybe that's just because the places where I use OSM are mapped 
wrong.


Using NE2's criteria, trunk is not really any different from a routing 
standpoint than primary. Using mine, trunk is something that would be 
primary if it weren't a physically better sort of road, so should be 
preferred over a nearby primary if the distances are similar. I'm 
harping on physical characteristics for this one particular tag because 
I think without that as a differentiator, it's basically useless fluff 
we shouldn't be using at all because primary already covers that usage.


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

Re: [Talk-us] US highway classification

[Anthony]

On Sat, May 28, 2011 at 10:40 AM, Mike N  wrote:
> On 5/28/2011 9:12 AM, Anthony wrote:
>>
>> Trunk has no meaning beyond "color the road the same color as other
>> things that are tagged trunk".
>
>  Even color is not defined - some trunks can be toll / not toll.
>
>  However, trunk *could* serve as a router hint that the road is a better
> selection than primary / secondary.

In my experience the difference between primary and trunk is generally
very minor, to the point where I'm not sure there'd be any advantage
at all in a router using it as a hint.

But maybe that's just because the places where I use OSM are mapped wrong.

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Sat, 28 May 2011 01:36:00 -0400, Nathan Edgars II wrote:



I mean best route, period. There's no diagonal Interstate there.


US-71 to I-44 to I-40 is faster. Not really a route I'd enjoy, but 
still faster.


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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/28/11 10:40 AM, Mike N wrote:

On 5/28/2011 9:12 AM, Anthony wrote:

Trunk has no meaning beyond "color the road the same color as other
things that are tagged trunk".


  Even color is not defined - some trunks can be toll / not toll.

  However, trunk *could* serve as a router hint that the road is a 
better selection than primary / secondary.

yes, it could, and that would be a valid reason for the distinction.

but for that, we need to use it consistently and there need to be
some focused meaning that usage.

richard


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

Re: [Talk-us] US highway classification

[Mike N]

On 5/28/2011 9:12 AM, Anthony wrote:

Trunk has no meaning beyond "color the road the same color as other
things that are tagged trunk".


  Even color is not defined - some trunks can be toll / not toll.

  However, trunk *could* serve as a router hint that the road is a 
better selection than primary / secondary.


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

Re: [Talk-us] US highway classification

[Anthony]

On Fri, May 27, 2011 at 10:04 AM, Nathan Mills  wrote:
> On Fri, 27 May 2011 09:26:41 -0400, Nathan Edgars II wrote:
>
>> No, trunk is also used for a major intercity highway that's not a
>> freeway. Take a look at the UK and their network of trunks.
>
> I'm sorry, I thought I posted to talk-us. My mistake. ;)
>
> Seems to me that trunk has no meaning if it is used in that way.

Trunk has no meaning beyond "color the road the same color as other
things that are tagged trunk".

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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 10:41 PM, Nathan Mills wrote:

Besides, if importance to the route network is the only consideration,
we ought not be using trunk at all or all US highways ought to be
classed as trunk.


Eh? A lot of U.S. Highways are no longer the most important highways, 
since they are paralleled by Interstates. Others never were the primary 
route (US 6 between Chicago and northwestern Pennsylvania, for example). 
In other cases, state-numbered highways are more major than roughly 
parallel U.S. Highways (example: SR 111 in eastern Tennessee is a 
four-lane Appalachian corridor, while US 127, a county to the east, is 
somewhat less).


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 9:51 PM, Nathan Mills wrote:

On Fri, 27 May 2011 12:17:53 -0400, Nathan Edgars II wrote:

Also, I don't know how major a road between Dumas, TX and Texline, TX
really is. If it weren't a US highway, I'd probably demote it all the
way to secondary.


It's on the National Highway System, meaning the FHWA considers it to
be a major road. It's probably the best route between Kansas City and
Albuquerque.


I'm going to assume you mean 'best non-Interstate route'. Most of it
isn't even four laned yet, although Texas has some of it under
construction. Same goes for the segment between Clayton, NM and I-25,
although there New Mexico is upgrading the road to four lane divided in
one whack. Which, as an aside, makes for one incredibly long
construction zone.


I mean best route, period. There's no diagonal Interstate there.



You seem to be arguing for using physical characteristics for trunk. But 
I (and some other mappers) don't see why this is necessary. Hence no 
consensus.


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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Fri, 27 May 2011 21:26:53 -0500, John F. Eldredge wrote:


I have driven on quite a few highways here in the USA that vary, mile
by mile, in the number of lanes, how well they are graded, whether or
not driveways connect directly to the highway, etc.  This usually
reflects their having been upgraded one piece at a time.  Sections
that pass through difficult terrain are often the last to be 
upgraded.

Of course, whether or not a local politician has friends or relatives
in the road-construction business makes a difference as well.

If you classify these highways according to their importance to the
transportation grid, then long sections, with variable physical
characteristics, will be classified the same.


Obviously some element of judgment is required no matter what. As you 
correctly point out, there is substantial variability in how roads are 
built in the US. If the substandard section is small, one could perhaps 
overlook it. If the upgraded section is small relative to the rest of 
the highway, perhaps it should not be used in determining the 
classification of the road. The intent here is not to classify solely on 
physical characteristics, but there is clearly a difference in the 
suitability of a road for long distance travel depending on whether it 
is 2 lane or 4 lane divided, and that should be reflected on the map, 
not just in tags.


I wouldn't downgrade a rural Interstate from motorway just because 
there are two driveways in three hundred miles that might see use three 
or four times a year. Nor would I upgrade a hundred miles of US highway 
from primary merely because of a mile of divided highway and one grade 
separated interchange.


Besides, if importance to the route network is the only consideration, 
we ought not be using trunk at all or all US highways ought to be 
classed as trunk. It seems obvious to me that neither can possibly be 
the sole consideration.


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

Re: [Talk-us] US highway classification

[John F. Eldredge]

athan Mills  wrote:

> On Fri, 27 May 2011 12:17:53 -0400, Nathan Edgars II wrote:
>
>>> The 'major intercity' road ought to be tagged as primary unless
>>> there's
>>> a specific reason to upgrade, IMO. That leaves the data more useful
>>> to
>>> end users.
>>
>> Actually that leaves it less useful for users in cities, as then
>> there are only two classifications for non-intercity highways,
>> secondary and tertiary.
>
> Uh, what are you on about? Motorway itself doesn't necessarily imply
>intercity or intracity, and neither do any of the other
>classifications.
>I can think of several intercity county roads that ought not qualify
>for
>anything beyond unclassified (they're old routes with several bypasses)
>
>and several intracity routes that definitely ought to be classified as
> trunk or motorway. It comes down to how the highway is built and what
> the highway is.
>
>>> Also, I don't know how major a road between Dumas, TX and Texline,
>>> TX
>>> really is. If it weren't a US highway, I'd probably demote it all
>>> the
>>> way to secondary.
>>
>> It's on the National Highway System, meaning the FHWA considers it to
>> be a major road. It's probably the best route between Kansas City and
>> Albuquerque.
>
> I'm going to assume you mean 'best non-Interstate route'. Most of it
> isn't even four laned yet, although Texas has some of it under
> construction. Same goes for the segment between Clayton, NM and I-25,
>although there New Mexico is upgrading the road to four lane divided in
>
> one whack. Which, as an aside, makes for one incredibly long
> construction zone.
>
> Talking solely about relatively rural areas, it seems to me that by
> default the best non-motorway route between two regionally important
>cities should be tagged primary unless there's a reason to upgrade it
>to
> trunk based on the physical characteristics of the road. To me, trunk
>implies a divided 4 lane at worst, or arguably including a true super
>2,
> of which I've seen a couple in Kansas (I think one of Oklahoma's
> turnpikes might also be a true super 2, but I haven't driven it
> personally). It just makes sense to me based on the way we build our
> roads here in the US.

I have driven on quite a few highways here in the USA that vary, mile by mile, 
in the number of lanes, how well they are graded, whether or not driveways 
connect directly to the highway, etc.  This usually reflects their having been 
upgraded one piece at a time.  Sections that pass through difficult terrain are 
often the last to be upgraded.  Of course, whether or not a local politician 
has friends or relatives in the road-construction business makes a difference 
as well.

If you classify these highways according to their importance to the 
transportation grid, then long sections, with variable physical 
characteristics, will be classified the same.

-- 
John F. Eldredge -- j...@jfeldredge.com
"Reserve your right to think, for even to think wrongly is better than not to 
think at all." -- Hypatia of Alexandria

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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Fri, 27 May 2011 12:17:53 -0400, Nathan Edgars II wrote:

The 'major intercity' road ought to be tagged as primary unless 
there's
a specific reason to upgrade, IMO. That leaves the data more useful 
to

end users.


Actually that leaves it less useful for users in cities, as then
there are only two classifications for non-intercity highways,
secondary and tertiary.


Uh, what are you on about? Motorway itself doesn't necessarily imply 
intercity or intracity, and neither do any of the other classifications. 
I can think of several intercity county roads that ought not qualify for 
anything beyond unclassified (they're old routes with several bypasses) 
and several intracity routes that definitely ought to be classified as 
trunk or motorway. It comes down to how the highway is built and what 
the highway is.


Also, I don't know how major a road between Dumas, TX and Texline, 
TX
really is. If it weren't a US highway, I'd probably demote it all 
the

way to secondary.


It's on the National Highway System, meaning the FHWA considers it to
be a major road. It's probably the best route between Kansas City and
Albuquerque.


I'm going to assume you mean 'best non-Interstate route'. Most of it 
isn't even four laned yet, although Texas has some of it under 
construction. Same goes for the segment between Clayton, NM and I-25, 
although there New Mexico is upgrading the road to four lane divided in 
one whack. Which, as an aside, makes for one incredibly long 
construction zone.


Talking solely about relatively rural areas, it seems to me that by 
default the best non-motorway route between two regionally important 
cities should be tagged primary unless there's a reason to upgrade it to 
trunk based on the physical characteristics of the road. To me, trunk 
implies a divided 4 lane at worst, or arguably including a true super 2, 
of which I've seen a couple in Kansas (I think one of Oklahoma's 
turnpikes might also be a true super 2, but I haven't driven it 
personally). It just makes sense to me based on the way we build our 
roads here in the US.


I maintain that tagging both two lane and four lane divided roads as 
trunk (not to mention the cases where it's used for a 
not-quite-a-motorway) makes the map much less useful for planning a 
route at a glance. Obviously, software can take the maxspeed and lanes 
tags into account when available but if I'm, for example, looking at 
some rendered tiles, that information is not available.


We already have four other tags to indicate importance in a route 
network, so I don't see the downside to limiting trunk to roads where 
the physical characteristics imply a higher classification, as we 
already do with motorway.


-wierdo

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

Re: [Talk-us] US highway classification

[Jason Straub]

Yeah, getting the definitions standardized has been thorny since day one.  I 
myself have had issues staying consistent from one road to the next.  There's 
just so much subtle differences between stretches of highways that make it 
difficult to say one way or the other.  Personally, I prefer the trunk being a 
non-limited access highway that has some type of divider.  Will use the Dumas 
to 
Texline stretch for example.  This stretch is the main trucking highway between 
Dallas and Denver, and is pretty much divided highway for all areas, excluding 
the stretches through the small cities that haven't been bypassed.  Otherwise 
this stretch might be considered a lonely stretch of little-traveled highway.  
This is where local knowledge comes in handy.  If standards could be agreed 
upon, then things would really fly.  Such are the perils of open sourcing...

Jason



On 5/27/2011 10:04 AM, Nathan Mills wrote:
> On Fri, 27 May 2011 09:26:41 -0400, Nathan Edgars II wrote:
>
>> No, trunk is also used for a major intercity highway that's not a
>> freeway. Take a look at the UK and their network of trunks.
>
> I'm sorry, I thought I posted to talk-us. My mistake. ;)
>
> Seems to me that trunk has no meaning if it is used in that way. In the
> UK, roads are classified based on the national government's
> classification of the roads (hence the confusing-to-us 'unclassified'
> tag. Since we don't have a single overarching national road network like
> that, I don't think that's a relevant model to use here.
How would you apply this argument to the use of primary to tertiary in 
the US?

> Generally
> speaking, I think any divided and controlled access highway probably
> ought to be tagged as a motorway barring specific local circumstances
> that cause it to deserve a demotion and any divided, but merely limited
> access and not fully controlled access, highway probably ought to be
> tagged as trunk barring specific local circumstances.
>
> The 'major intercity' road ought to be tagged as primary unless there's
> a specific reason to upgrade, IMO. That leaves the data more useful to
> end users.

Actually that leaves it less useful for users in cities, as then there 
are only two classifications for non-intercity highways, secondary and 
tertiary.
>
> Also, I don't know how major a road between Dumas, TX and Texline, TX
> really is. If it weren't a US highway, I'd probably demote it all the
> way to secondary.

It's on the National Highway System, meaning the FHWA considers it to be 
a major road. It's probably the best route between Kansas City and 
Albuquerque.

Also note the proposed translation on 
http://wiki.openstreetmap.org/wiki/Highway_Functional_Classification_System 
, used by at least one mapper in Kansas. Principal arterials range from 
expressways to major two-lane intercity highways.



--

Message: 3
Date: Fri, 27 May 2011 12:21:09 -0400
From: Nathan Edgars II 
To: talk-us@openstreetmap.org
Subject: Re: [Talk-us] US highway classification
Message-ID: <4ddfcf75.9080...@gmail.com>
Content-Type: text/plain; charset=windows-1252; format=flowed

On 5/27/2011 12:00 PM, Alex Mauer wrote:
> On 05/27/2011 09:06 AM, Richard Welty wrote:
>> if you peruse the wiki, and make a reasonably through search
>> for definitions of trunk in the US, you will find an extensive
>> complex of contradictions and inconsistencies.
>
> Maybe someone should find all these and bring it up on the list so that
> a definition can be determined and the inconsistencies can be fixed?

I have tried in the past. The problem is that nobody can agree on which 
definition to use (even in the trunks must have four lanes camp, there's 
the Texas style, where everything with four lanes and a median is trunk, 
and the expressway style, where only limited-access surface expressways 
are trunk). Hell, even motorway is controversial (Seattle's Alaskan Way 
Viaduct is marked as trunk when it should be motorway).

I'm all for consistency, even if it's not the way I currently tag. But 
this doesn't seem to be an attainable goal.

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

Re: [Talk-us] US highway classification

[Greg Troxel]

Richard Welty  writes:

> On 5/27/11 9:26 AM, Nathan Edgars II wrote:
>> On 5/27/2011 12:32 AM, Nathan Mills wrote:
>>> Would I be correct in stating that tagging an undivided 2 lane (one lane
>>> in each direction) highways would be improper, even if a state calls the
>>> highway a "trunk" for planning purposes? Especially if it's in the
>>> middle of a town with a low speed limit. I understood trunk to be
>>> divided and limited access (but not fully grade-separated).
>>
>> No, trunk is also used for a major intercity highway that's not a
>> freeway. Take a look at the UK and their network of trunks.
>>
> and we have no functioning definition of what constitutes a major intercity
> highway in the US.
>
> which is to day, i don't agree with many of your upgrades, but i've chosen
> not to engage in an edit war in the map nor have i chosen to engage in
> flame wars on this list about it, but there is no consensus on what
> constitutes
> a trunk in the US, and i believe your position is an outlier.

Emphatically seconded.

I think it's clear that most of us think trunk is something that has
some aspects of divided, particularly high speed, limited access above
and beyond a normal two lane US highway.

I find the trunk designation useful in that it tells the map viewer that
the road is partway to feeling like an Interstate compared to feeling
like a US highway.   Marking roads as trunk when they aren't physically
superior isn't helpful.


pgpFMGOS1f1nR.pgp
Description: PGP signature
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 12:00 PM, Alex Mauer wrote:

On 05/27/2011 09:06 AM, Richard Welty wrote:

if you peruse the wiki, and make a reasonably through search
for definitions of trunk in the US, you will find an extensive
complex of contradictions and inconsistencies.


Maybe someone should find all these and bring it up on the list so that
a definition can be determined and the inconsistencies can be fixed?


I have tried in the past. The problem is that nobody can agree on which 
definition to use (even in the trunks must have four lanes camp, there's 
the Texas style, where everything with four lanes and a median is trunk, 
and the expressway style, where only limited-access surface expressways 
are trunk). Hell, even motorway is controversial (Seattle's Alaskan Way 
Viaduct is marked as trunk when it should be motorway).


I'm all for consistency, even if it's not the way I currently tag. But 
this doesn't seem to be an attainable goal.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 10:04 AM, Nathan Mills wrote:

On Fri, 27 May 2011 09:26:41 -0400, Nathan Edgars II wrote:


No, trunk is also used for a major intercity highway that's not a
freeway. Take a look at the UK and their network of trunks.


I'm sorry, I thought I posted to talk-us. My mistake. ;)

Seems to me that trunk has no meaning if it is used in that way. In the
UK, roads are classified based on the national government's
classification of the roads (hence the confusing-to-us 'unclassified'
tag. Since we don't have a single overarching national road network like
that, I don't think that's a relevant model to use here.
How would you apply this argument to the use of primary to tertiary in 
the US?


> Generally

speaking, I think any divided and controlled access highway probably
ought to be tagged as a motorway barring specific local circumstances
that cause it to deserve a demotion and any divided, but merely limited
access and not fully controlled access, highway probably ought to be
tagged as trunk barring specific local circumstances.

The 'major intercity' road ought to be tagged as primary unless there's
a specific reason to upgrade, IMO. That leaves the data more useful to
end users.


Actually that leaves it less useful for users in cities, as then there 
are only two classifications for non-intercity highways, secondary and 
tertiary.


Also, I don't know how major a road between Dumas, TX and Texline, TX
really is. If it weren't a US highway, I'd probably demote it all the
way to secondary.


It's on the National Highway System, meaning the FHWA considers it to be 
a major road. It's probably the best route between Kansas City and 
Albuquerque.


Also note the proposed translation on 
http://wiki.openstreetmap.org/wiki/Highway_Functional_Classification_System 
, used by at least one mapper in Kansas. Principal arterials range from 
expressways to major two-lane intercity highways.


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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/27/11 12:00 PM, Alex Mauer wrote:

On 05/27/2011 09:06 AM, Richard Welty wrote:

if you peruse the wiki, and make a reasonably through search
for definitions of trunk in the US, you will find an extensive
complex of contradictions and inconsistencies.

Maybe someone should find all these and bring it up on the list so that
a definition can be determined and the inconsistencies can be fixed?

Just saying “the definition is inconsistent so I’ll just use my own
interpretation” isn’t very constructive.

as it happens, i did this a while back. there may have been edits since
i collected this stuff, but i included the URLs where the text came from
for a reason.

Definitions found in the Wiki:

Generic definition from
 http://wiki.openstreetmap.org/wiki/Map_Features#Highway

  Important roads that aren't motorways. Typically maintained by central,
  not local government. Need not necessarily be a divided highway. In the
  UK, all green signed A roads are, in OSM, classed as 'trunk'.


From http://wiki.openstreetmap.org/wiki/Trunk#International_equivalence


  Divided highway without ramps.
  Higher speeds (65+mph). This includes some US highways and some state
  highways.

From 

http://wiki.openstreetmap.org/wiki/United_States_roads_tagging#Trunk_tag
(Interstate Highway section of page)

  Trunk tag

  NOTE: The definition below is not commonly used; see the talk page.
  Many people use trunk to mean "expressway"-grade arterials with at-grade
  intersections, major non-motorway intercity highways, or both.

  Most controlled-access highways without adequate speed or travel lanes
  or with obstructions should be designated highway=trunk. This
  designation applies, for example, to the two-lane Interstate 93 in
  northern New Hampshire. highway=trunk should apply to any segment,
  travel on which typically implies or necessitates clearing the
  obstruction. Any ramps onto or from a trunk highway get
  highway=trunk_link, even if they otherwise qualify for
  highway=motorway_link. Ramps leading into or from weigh stations,
  inspection booths, welcome centers, rest areas, and similar diversions
  accessible only from a trunk or motorway highway also carry
  highway=trunk_link.

  Trunk highways include controlled-access highways that lie within
  military bases; contain draw bridges, toll booths, or other obstructions;
  have a speed limit less than 50 miles per hour; or have only one lane in
  each direction, whether divided or  otherwise (sometimes called a
  "super-two" freeway). The designations highway=trunk and
  highway=trunk_link apply to all toll roads.

From
http://wiki.openstreetmap.org/wiki/United_States_roads_tagging#Motorway_and_Trunk_tags
(United States Highway section of page)

Motorway and Trunk tags

  If any segment of a U.S. highway or any other road merits
  highway=motorway or highway=trunk according to the criteria
  heretofore described, it should be so designated.

From the talk side of US tagging (note that there is extensive 
discussion which

i have not copied into this document)
http://wiki.openstreetmap.org/wiki/Talk:United_States_roads_tagging#Trunk

  Trunk

  Unless part of the "Interstate system"

* An Arterial Divided highway that is partially but not entirely 
grade separated.


From http://wiki.openstreetmap.org/wiki/Roads#USA

  US Highways should be tagged with highway=primary. State Highways and
  County Highways should be tagged with highway=secondary. Any of these
  which is a divided highway with high speeds (65mph+) and intersections
  with other roads, and legal for bicycles and motorbikes to use, should be
  tagged with highway=trunk.



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

Re: [Talk-us] US highway classification

[Alex Mauer]

On 05/27/2011 09:06 AM, Richard Welty wrote:
> if you peruse the wiki, and make a reasonably through search
> for definitions of trunk in the US, you will find an extensive
> complex of contradictions and inconsistencies.

Maybe someone should find all these and bring it up on the list so that
a definition can be determined and the inconsistencies can be fixed?

Just saying “the definition is inconsistent so I’ll just use my own
interpretation” isn’t very constructive.

—Alex Mauer “hawke”


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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/27/11 9:47 AM, Nathan Edgars II wrote:

On 5/27/2011 9:34 AM, Richard Welty wrote:
which is to day, i don't agree with many of your upgrades, but i've 
chosen

not to engage in an edit war in the map nor have i chosen to engage in
flame wars on this list about it, but there is no consensus on what
constitutes
a trunk in the US, and i believe your position is an outlier.


I've tagged based on what practices I've seen applied to the map. For 
example, when I joined, several two-lane highways across Nevada were 
trunk. New Jersey has had a number of two-lane trunks for years.

if you peruse the wiki, and make a reasonably through search
for definitions of trunk in the US, you will find an extensive
complex of contradictions and inconsistencies. it is not actually
possible to use trunk as a classification in the US without
violating one or another constraint in the wiki as it currently
stands.

the fact that there are trunks of this that or the other type
scattered about the map of the US doesn't mean there is
consensus, it merely means that some mappers happen to
use a particular definition of trunk.

like i say, no consensus. my reading of the tenor of past discussions
of the list suggests that you are classifying highways as trunk that most
would probably classify as primary (you've also upgraded a bunch of
secondaries in upstate NY to primary where i have local knowledge of
the roads in question and don't really agree with the upgrade.)

richard



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

Re: [Talk-us] US highway classification

[Nathan Mills]

On Fri, 27 May 2011 09:26:41 -0400, Nathan Edgars II wrote:


No, trunk is also used for a major intercity highway that's not a
freeway. Take a look at the UK and their network of trunks.


I'm sorry, I thought I posted to talk-us. My mistake. ;)

Seems to me that trunk has no meaning if it is used in that way. In the 
UK, roads are classified based on the national government's 
classification of the roads (hence the confusing-to-us 'unclassified' 
tag. Since we don't have a single overarching national road network like 
that, I don't think that's a relevant model to use here. Generally 
speaking, I think any divided and controlled access highway probably 
ought to be tagged as a motorway barring specific local circumstances 
that cause it to deserve a demotion and any divided, but merely limited 
access and not fully controlled access, highway probably ought to be 
tagged as trunk barring specific local circumstances.


The 'major intercity' road ought to be tagged as primary unless there's 
a specific reason to upgrade, IMO. That leaves the data more useful to 
end users.


Also, I don't know how major a road between Dumas, TX and Texline, TX 
really is. If it weren't a US highway, I'd probably demote it all the 
way to secondary.


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 9:34 AM, Richard Welty wrote:

On 5/27/11 9:26 AM, Nathan Edgars II wrote:

On 5/27/2011 12:32 AM, Nathan Mills wrote:

Would I be correct in stating that tagging an undivided 2 lane (one lane
in each direction) highways would be improper, even if a state calls the
highway a "trunk" for planning purposes? Especially if it's in the
middle of a town with a low speed limit. I understood trunk to be
divided and limited access (but not fully grade-separated).


No, trunk is also used for a major intercity highway that's not a
freeway. Take a look at the UK and their network of trunks.


and we have no functioning definition of what constitutes a major intercity
highway in the US.


There's no functioning definition of any of the classifications below 
motorway.


which is to day, i don't agree with many of your upgrades, but i've chosen
not to engage in an edit war in the map nor have i chosen to engage in
flame wars on this list about it, but there is no consensus on what
constitutes
a trunk in the US, and i believe your position is an outlier.


I've tagged based on what practices I've seen applied to the map. For 
example, when I joined, several two-lane highways across Nevada were 
trunk. New Jersey has had a number of two-lane trunks for years.


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

Re: [Talk-us] US highway classification

[Richard Welty]

On 5/27/11 9:26 AM, Nathan Edgars II wrote:

On 5/27/2011 12:32 AM, Nathan Mills wrote:

Would I be correct in stating that tagging an undivided 2 lane (one lane
in each direction) highways would be improper, even if a state calls the
highway a "trunk" for planning purposes? Especially if it's in the
middle of a town with a low speed limit. I understood trunk to be
divided and limited access (but not fully grade-separated).


No, trunk is also used for a major intercity highway that's not a 
freeway. Take a look at the UK and their network of trunks.



and we have no functioning definition of what constitutes a major intercity
highway in the US.

which is to day, i don't agree with many of your upgrades, but i've chosen
not to engage in an edit war in the map nor have i chosen to engage in
flame wars on this list about it, but there is no consensus on what 
constitutes

a trunk in the US, and i believe your position is an outlier.

richard


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

Re: [Talk-us] US highway classification

[Nathan Edgars II]

On 5/27/2011 12:32 AM, Nathan Mills wrote:

Would I be correct in stating that tagging an undivided 2 lane (one lane
in each direction) highways would be improper, even if a state calls the
highway a "trunk" for planning purposes? Especially if it's in the
middle of a town with a low speed limit. I understood trunk to be
divided and limited access (but not fully grade-separated).


No, trunk is also used for a major intercity highway that's not a 
freeway. Take a look at the UK and their network of trunks.


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

Re: [Talk-us] US highway classification

[Mike N]

On 5/27/2011 12:32 AM, Nathan Mills wrote:

Would I be correct in stating that tagging an undivided 2 lane (one lane
in each direction) highways would be improper, even if a state calls the
highway a "trunk" for planning purposes?


I agree for this case, there is no established convention that can 
consider this a trunk.


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

Re: [Talk-us] US Highway Classification (Was: directions of ways in MassGIS data)

[Alan Brown]

There was a discussion on this list ("Road classification") a short while ago about the "Highway" tag that should be revisited. Someone posted a classification system back in December that made a lot of sense, and is more in sync with what I know commercial data providers.  Basically, it would be treating the "highway" tag as providing a functional road classification, rather than (primarily) a legal classification or a physical description.  In a nutshell, "Primary" = Road to take you between cities, "Secondary" = Road to take you across cities, "Tertiary" = Road to take you through neighborhoods.  It ends up roughly correlating with physical or legal designations - but not exactly.  It's marking the relative importance of a particular road.  (Motorways and Trunk Roads are
 exceptions to this, as they do have specific physical descriptions.) Because of this, I quite disagree with the statement that the tag should be used as a general description of the physical attributes of a road; there are physical description tags to serve that purpose.  There are many cases where a two lane highway is by far the most important road in a region (Transcanadian Highway, US 50 in Nevada), while in urban areas like San Jose, you have relatively unimportant roads with two or three lanes in each direction for short distances.  Yes, there is definitely subjective judgement here, but it's the sort of subjective cartographers make on a regular basis, and it is very useful.  You want a good balance between primary, secondary, and tertiary roads.  -AlanHere is map...@att.net's old post:In OSM language the Highway feature is used to designate what we call roads. 
 
A motorway is a four+ lane, limited access, grade separated freeway. These can include Interstates, US Highways, State Highways, County Highways
or even Farm to Market Roads if they meet certain criteria. These
criteria are limited access,the use entrance/exit ramps to access the
freeway. Intersections with other roads are at grade seperated
crossings or ramps. A grade separated crossing means one road goes over
or under the other. (ie. over/underpass) When Motorways meet other
motorways they generally use ramps that are classified as Motorway
Link. These motorways usually connect to other cities or move the
traffic around and through a city. Limited access ring roads usually
fall in this feature class also. 
 
A trunk is what a motorway becomes when it loses one of it's
criteria. This usually occurs to US, State, County highways as they
move outside the urban areas. Intersections with other roads can occur
at grade and/or when ramps are no longer needed to access the road.
Usually they remain 4+ lanes and may or may not be divided by a
physical median. 
 
A primary road can be a US, State, County Highway or other road
that connects two cities or moves traffic from one part of the city to
the other. These are the highways that become Main St when they go
through a small rural town. They will have traffic signals
when they reach more densely populated areas. These are the roads you
jump on when the freeway has an accident and you don't want to sit and
wait it out. 
 
A secondary road moves traffic within a city. It would service only a certain area within a city. 
 
A tertiary road connects the residential roads to the higher classes: motorway, trunk, primary or secondary. 
 
I hope this clarifys things for some users. I know it's not going to please those who have already used other classification schemes.
 
thanks,
pete


From: David Lynch To: Talk-us Sent: Tuesday, February 3, 2009 8:49:26 AMSubject: Re: [Talk-us] US Highway Classification (Was: directions of ways in MassGIS data)
What I've tended to do in my part of Texas is:Motorway - two or more consecutive intersections with grade separation and no driveways, or any interstate (some very rural locations do have the occasional turn off directly from the main travel lanes)
Trunk - US highways without any other reason to classify them up or down and high speed divided roads with relatively few crossings and an occasional grade separationPrimary - State highways, US highways running near and parallel to a motorway, and wide, heavily-traveled urban/suburban streets
Secondary - Other state highways (farm-to-market roads, loops, spurs) and major city streetsTertiary - Residential collector streets, some rural roads that are in good condition and connect to classified highways, and service roads running parallel to a motorway.
On Tue, Feb 3, 2009 at 09:44, Zeke Farwell <ezeki...@gmail.com> wrote:
We need a detailed road classification chart on the US wiki page to straighten all of this out !  Maybe I'll find the time to get working on that one of these days...  I think I'm also having trouble making the OSM road classification system do everything I want it to.  

"the consensus is that the highway ta

Re: [Talk-us] US Highway Classification (Was: directions of ways in MassGIS data)

[David Lynch]

What I've tended to do in my part of Texas is:

Motorway - two or more consecutive intersections with grade separation and
no driveways, or any interstate (some very rural locations do have the
occasional turn off directly from the main travel lanes)
Trunk - US highways without any other reason to classify them up or down and
high speed divided roads with relatively few crossings and an occasional
grade separation
Primary - State highways, US highways running near and parallel to a
motorway, and wide, heavily-traveled urban/suburban streets
Secondary - Other state highways (farm-to-market roads, loops, spurs) and
major city streets
Tertiary - Residential collector streets, some rural roads that are in good
condition and connect to classified highways, and service roads running
parallel to a motorway.

On Tue, Feb 3, 2009 at 09:44, Zeke Farwell  wrote:

> We need a detailed road classification chart on the US wiki page to
> straighten all of this out [?]!  Maybe I'll find the time to get working on
> that one of these days...  I think I'm also having trouble making the OSM
> road classification system do everything I want it to.
>
> "the consensus is that the highway tag is for making a general description
> of the *physical* attributes of a highway. This gives the user of a
> highway more useful information than a legal classification."  -  Key:
> Highway (OSM Wiki)  
>
> This is true for high zoom levels, but at low zoom levels this can be too
> much information.  When I am looking at a map of an entire state, I want to
> see what the main routes are through that state, not every important road.
>
> In Vermont some important 2 lane north/south routes are VT 100, VT 116, US
> 7, and VT 22A.  Physically they are pretty similar, and so by OSM standards
> they should all be primary.  In my opinion though, at low zoom levels US 7
> should stand out from the others because it connects larger towns, has a
> higher legal designation, and it stands out on other maps.
>
> I suppose this is really an issue for renderers, but we do need some sort
> of tagging for the renderers to use.  Maybe once route relations become more
> prevalent renderers could emphasise different higher level route networks
> more, regardless of the physical highway tag.  Any thoughts?
>
> Zeke Farwell
> Burlington, VT
>
>
>
> On Mon, Feb 2, 2009 at 10:20 PM, Greg Troxel  wrote:
>
>> (replying to Zeke and Chris both)
>>
>> I agree that if there is only 1 mile of motorway class road among
>> trunk-class road that tagging it motorway isn't useful.
>>
>> The parts of Route 2 that I was thinking of tagging as motorway are
>> physically indistinguishable from an interstate, and at least 10 miles
>> long.  I am not 100% clear that pedestrians/bicycles are banned, but I
>> think someone would call the police if either were on the road - I never
>> see them when driving on 2.`
>>
>> As for the trunk designation, I find that a bit ocnfusing, but my
>> impression is that it is a road that is somewhat more than a regular US
>> highway physically, but not all the way to interstate.  An example would
>> be a road with 2 lanes in each direction and jersey barriers, but
>> same-grade junctions with lights every few miles.  Rt 2 is like this
>> between 128 and Tracey's corner (1st intersection west) and really all
>> the way to South Acton where it is motorway again until about 7 miles E
>> of Orange and then it's back to trunk (1 lane each way only, but with
>> exits).  Further west it is just a primary road, but gradually being
>> made more like trunk.
>>
>> I don't think being the only main road should qualify a way as trunk; it
>> seems being important is only enough to get a road to primary.  But some
>> degree of limited access and being divided would be enough for trunk.
>> So I'd leave rt 2 as motorway/trunk mixed.  I've driven rt 7 from
>> Bennington to Burlington, and it didn't in general feel like "almost an
>> interstate but not quite" - it felt on balance more like primary.
>>
>>
>>
>>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>
>


-- 
David J. Lynch
djly...@gmail.com
<<338.png>>___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US Highway Classification (Was: directions of ways in MassGIS data)

[Zeke Farwell]

We need a detailed road classification chart on the US wiki page to
straighten all of this out [?]!  Maybe I'll find the time to get working on
that one of these days...  I think I'm also having trouble making the OSM
road classification system do everything I want it to.

"the consensus is that the highway tag is for making a general description
of the *physical* attributes of a highway. This gives the user of a highway
more useful information than a legal classification."  -  Key: Highway (OSM
Wiki)  

This is true for high zoom levels, but at low zoom levels this can be too
much information.  When I am looking at a map of an entire state, I want to
see what the main routes are through that state, not every important road.

In Vermont some important 2 lane north/south routes are VT 100, VT 116, US
7, and VT 22A.  Physically they are pretty similar, and so by OSM standards
they should all be primary.  In my opinion though, at low zoom levels US 7
should stand out from the others because it connects larger towns, has a
higher legal designation, and it stands out on other maps.

I suppose this is really an issue for renderers, but we do need some sort of
tagging for the renderers to use.  Maybe once route relations become more
prevalent renderers could emphasise different higher level route networks
more, regardless of the physical highway tag.  Any thoughts?

Zeke Farwell
Burlington, VT



On Mon, Feb 2, 2009 at 10:20 PM, Greg Troxel  wrote:

> (replying to Zeke and Chris both)
>
> I agree that if there is only 1 mile of motorway class road among
> trunk-class road that tagging it motorway isn't useful.
>
> The parts of Route 2 that I was thinking of tagging as motorway are
> physically indistinguishable from an interstate, and at least 10 miles
> long.  I am not 100% clear that pedestrians/bicycles are banned, but I
> think someone would call the police if either were on the road - I never
> see them when driving on 2.`
>
> As for the trunk designation, I find that a bit ocnfusing, but my
> impression is that it is a road that is somewhat more than a regular US
> highway physically, but not all the way to interstate.  An example would
> be a road with 2 lanes in each direction and jersey barriers, but
> same-grade junctions with lights every few miles.  Rt 2 is like this
> between 128 and Tracey's corner (1st intersection west) and really all
> the way to South Acton where it is motorway again until about 7 miles E
> of Orange and then it's back to trunk (1 lane each way only, but with
> exits).  Further west it is just a primary road, but gradually being
> made more like trunk.
>
> I don't think being the only main road should qualify a way as trunk; it
> seems being important is only enough to get a road to primary.  But some
> degree of limited access and being divided would be enough for trunk.
> So I'd leave rt 2 as motorway/trunk mixed.  I've driven rt 7 from
> Bennington to Burlington, and it didn't in general feel like "almost an
> interstate but not quite" - it felt on balance more like primary.
>
>
>
>
<<338.png>>___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us