Re: [Tagging] lanes tag dispute

2011-09-19 Thread Anthony
2011/9/19 Kytömaa Lauri lauri.kyto...@aalto.fi:
 If there is one globally accepted definition of what a lane is, it's the
 one in the UN Vienna Convention on road signs and signals:

 Lane means any one of the longitudinal strips into which the carriageway
 is divisible, whether or not defined by longitudinal road markings, which
 is wide enough for one moving line of motor vehicles other than motor
 cycles;

So, this includes parking lanes?  Or am I misreading it?

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


Re: [Tagging] lanes tag dispute

2011-09-19 Thread Anthony
On Mon, Sep 19, 2011 at 5:41 AM, Martin Koppenhoefer
dieterdre...@gmail.com wrote:
 2011/9/19 Kytömaa Lauri lauri.kyto...@aalto.fi:
 Lane means any one of the longitudinal strips into which the carriageway
 is divisible, whether or not defined by longitudinal road markings, which
 is wide enough for one moving line of motor vehicles other than motor
 cycles;


 OK, this would exclude cyclelanes. To me this definition seems OK

Seems okay to me too, pending clarification about parking lanes, and
with the caveat that this definition only applies to ways which
actually have motor vehicle lanes (i.e. it wouldn't apply to a laned
cycleway).

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Tobias Knerr
Nathan Edgars II wrote:
 Currently user Alv is trying to redefine the lanes tag to say that it
 must include all turn lanes:
 http://wiki.openstreetmap.org/w/index.php?title=Key:lanesaction=history

Since 2008, the lane tag has been defined as the _physical_ number of
lanes. In other words, it doesn't matter what these lanes are intended
for: Whether they are, for example, reserved to high occupancy vehicles,
or whether they are turn lanes. At least that's how I understand the
wiki page.

Alv's edits are consistent with this definition, so I do not agree that
removing your addition to the page is a redefinition of the lanes tag.

Tobias Knerr

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Erik Johansson
On Sun, Sep 18, 2011 at 1:47 PM, Nathan Edgars II nerou...@gmail.com wrote:
 Currently user Alv is trying to redefine the lanes tag to say that it must
 include all turn lanes:
 http://wiki.openstreetmap.org/w/index.php?title=Key:lanesaction=history

I've had this discussion before 2007 and I've always said and heard
that lanes is always all the lanes on the road. This very easy
definition goes back to the first version of the key:lanes page. See:

http://wiki.openstreetmap.org/w/index.php?title=Key:lanesoldid=85670

So the current wikipage is wrong and too complex.
-- 
/emj

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Martin Koppenhoefer
2011/9/18 Erik Johansson erjo...@gmail.com:
 On Sun, Sep 18, 2011 at 1:47 PM, Nathan Edgars II nerou...@gmail.com wrote:
 Currently user Alv is trying to redefine the lanes tag to say that it must
 include all turn lanes:
 http://wiki.openstreetmap.org/w/index.php?title=Key:lanesaction=history

 I've had this discussion before 2007 and I've always said and heard
 that lanes is always all the lanes on the road. This very easy
 definition goes back to the first version of the key:lanes page. See:

 http://wiki.openstreetmap.org/w/index.php?title=Key:lanesoldid=85670


+1, lanes from the very beginning were to indicate all lanes of the
road. Opposed to this was the common practice, not to split an highway
because of a short lane for turns (otherwise we would for example have
had to split a motorway at every exit which I never saw in the actual
data).

cheers,
Martin

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 8:33 AM, Martin Koppenhoefer wrote:

+1, lanes from the very beginning were to indicate all lanes of the
road. Opposed to this was the common practice, not to split an highway
because of a short lane for turns (otherwise we would for example have
had to split a motorway at every exit which I never saw in the actual
data).


So it's down to the old prescriptive vs. descriptive debate then.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Erik Johansson
On Sun, Sep 18, 2011 at 2:36 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 8:33 AM, Martin Koppenhoefer wrote:

 +1, lanes from the very beginning were to indicate all lanes of the
 road. Opposed to this was the common practice, not to split an highway
 because of a short lane for turns (otherwise we would for example have
 had to split a motorway at every exit which I never saw in the actual
 data).

 So it's down to the old prescriptive vs. descriptive debate then.

The current lanes tag works pretty well for me as a pedestrian, it
tells me how many lanes there is on a road.  I agree that it gives too
little information to be useful for routers etc, but that's because
the generic name. Please write something more descriptive to describe
the problem and how you think it should be solved. These single
sentence responses do not really help.

Btw the current wiki page seems to say that you should count cycle
lanes in the lanes tag, something I don't really agree with.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Martin Koppenhoefer
2011/9/18 Erik Johansson erjo...@gmail.com:
 Btw the current wiki page seems to say that you should count cycle
 lanes in the lanes tag, something I don't really agree with.


+1, maybe we should specify that lanes is about car-lanes (restricted
lanes like bus-lanes or cycle-lanes should IMHO be tagged with
separate tags.)

cheers,
Martin

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 9:12 AM, Erik Johansson wrote:

The current lanes tag works pretty well for me as a pedestrian, it
tells me how many lanes there is on a road.
This could be more usefully placed in the crossing node. Outside a 
crosswalk, it's much more useful to know if there's a shelter (raised 
median or paved and striped) in the middle of the road.


Here's an example of what tagging every lane would lead to: 
http://wiki.openstreetmap.org/wiki/File:Silly_lane_count.jpg


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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Dave F.

On 18/09/2011 14:21, Martin Koppenhoefer wrote:

2011/9/18 Erik Johanssonerjo...@gmail.com:

Btw the current wiki page seems to say that you should count cycle
lanes in the lanes tag, something I don't really agree with.


+1, maybe we should specify that lanes is about car-lanes (restricted
lanes like bus-lanes or cycle-lanes should IMHO be tagged with
separate tags.)


I agree.

Dave F.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Dave F.

On 18/09/2011 16:20, Nathan Edgars II wrote:

On 9/18/2011 9:12 AM, Erik Johansson wrote:

The current lanes tag works pretty well for me as a pedestrian, it
tells me how many lanes there is on a road.
This could be more usefully placed in the crossing node. Outside a 
crosswalk, it's much more useful to know if there's a shelter (raised 
median or paved and striped) in the middle of the road.


I agree it's more useful there, but please remember that jaywalking 
isn't an offence in many countries.




Here's an example of what tagging every lane would lead to: 
http://wiki.openstreetmap.org/wiki/File:Silly_lane_count.jpg 


I see nothing intrinsically wrong with that. Lots of splitting ways  
tagging admittedly, but that's a problem with OSM  its tagging process.


Dave F.




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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 1:12 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 12:59 PM, Nathan Edgars IInerou...@gmail.com  wrote:

I would use lanes=2 there, since that's how many through lanes there are
before the turn lanes begin.


And you think any other number of lanes would be *wrong*?


No. What I don't agree with is that lanes=2 is missing data or an 
interim solution. It is simply another way of tagging.


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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 1:18 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 1:12 PM, Anthonyo...@inbox.org  wrote:

Yes, mapping every lane as a separate way would be even better, but
let's go one step at a time.


Actually, I take back that last sentence:
http://www.pocketgpsworld.com/reviews/tomtom-v8/v8_lane_assist_northwest2_speed_alert.jpg

Do you think Garmin is mapping every lane as a separate way here?  I
don't think they are, or that they should.  I think if I were going to
do it, I'd map this as one way up to the theoretical gore point, and
then as two ways after the theoretical gore point.  I wouldn't use
five ways.


It's simple on a motorway. But, for a simple example on a surface road, 
how do you know where the straight lanes go if there are 3 lanes after 
an intersection but 2 before?


And actually, even on a motorway, what if it's 4 lanes to a split of 2 
and 3? The assumption is that the second lane from the left can be used 
to exit, but it's possible that only the leftmost lane is available, and 
a second lane begins on the ramp right at the gore. Or perhaps there's a 
short segment of 5 lanes before the split - how do you know which side 
the new lane forms on?


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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 1:17 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 1:12 PM, Anthony wrote:

 On Sun, Sep 18, 2011 at 12:59 PM, Nathan Edgars IInerou...@gmail.com
  wrote:

 I would use lanes=2 there, since that's how many through lanes there are
 before the turn lanes begin.

 And you think any other number of lanes would be *wrong*?

 No. What I don't agree with is that lanes=2 is missing data or an interim
 solution. It is simply another way of tagging.

I couldn't disagree with that sentiment more.  Tagging the right
number of lanes is clearly better than tagging the number of lanes
that there used to be.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 1:21 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 1:18 PM, Anthony wrote:
 http://www.pocketgpsworld.com/reviews/tomtom-v8/v8_lane_assist_northwest2_speed_alert.jpg

 Do you think Garmin is mapping every lane as a separate way here?  I
 don't think they are, or that they should.  I think if I were going to
 do it, I'd map this as one way up to the theoretical gore point, and
 then as two ways after the theoretical gore point.  I wouldn't use
 five ways.

 It's simple on a motorway. But, for a simple example on a surface road, how
 do you know where the straight lanes go if there are 3 lanes after an
 intersection but 2 before?

Depends on what data you have.  If you have lane width information,
then it's easy to figure this out.  If you don't, then you can't
figure this out.  If all lanes are parallel, you certainly don't need
a way for every lane, though.

But why is this important in the first place?

 And actually, even on a motorway, what if it's 4 lanes to a split of 2 and
 3? The assumption is that the second lane from the left can be used to exit,
 but it's possible that only the leftmost lane is available, and a second
 lane begins on the ramp right at the gore. Or perhaps there's a short
 segment of 5 lanes before the split - how do you know which side the new
 lane forms on?

This can be determined by the geometry of the ways, which are mapped
at the center.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 1:34 PM, Anthony o...@inbox.org wrote:
 On Sun, Sep 18, 2011 at 1:21 PM, Nathan Edgars II nerou...@gmail.com wrote:
 It's simple on a motorway. But, for a simple example on a surface road, how
 do you know where the straight lanes go if there are 3 lanes after an
 intersection but 2 before?

 Depends on what data you have.  If you have lane width information,
 then it's easy to figure this out.  If you don't, then you can't
 figure this out.

Nevermind.  Yes you can.  If the center of the 3 lane road is to the
left of the center of the 2 lane road, then the lane was added on the
left.  If the center of the 3 lane road is to the right of the center
of the 2 lane road, then the lane was added on the right.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 1:34 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 1:21 PM, Nathan Edgars IInerou...@gmail.com  wrote:

On 9/18/2011 1:18 PM, Anthony wrote:

http://www.pocketgpsworld.com/reviews/tomtom-v8/v8_lane_assist_northwest2_speed_alert.jpg

Do you think Garmin is mapping every lane as a separate way here?  I
don't think they are, or that they should.  I think if I were going to
do it, I'd map this as one way up to the theoretical gore point, and
then as two ways after the theoretical gore point.  I wouldn't use
five ways.


It's simple on a motorway. But, for a simple example on a surface road, how
do you know where the straight lanes go if there are 3 lanes after an
intersection but 2 before?


Depends on what data you have.  If you have lane width information,
then it's easy to figure this out.  If you don't, then you can't
figure this out.  If all lanes are parallel, you certainly don't need
a way for every lane, though.

But why is this important in the first place?


If you have two intersections in quick succession, and are turning at 
the second, you want to know which lane to be in at the first. That is, 
given that we want this sort of thing in the first place.



And actually, even on a motorway, what if it's 4 lanes to a split of 2 and
3? The assumption is that the second lane from the left can be used to exit,
but it's possible that only the leftmost lane is available, and a second
lane begins on the ramp right at the gore. Or perhaps there's a short
segment of 5 lanes before the split - how do you know which side the new
lane forms on?


This can be determined by the geometry of the ways, which are mapped
at the center.


No they're not.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 1:39 PM, Nathan Edgars II wrote:

On 9/18/2011 1:38 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 1:34 PM, Anthonyo...@inbox.org wrote:

On Sun, Sep 18, 2011 at 1:21 PM, Nathan Edgars IInerou...@gmail.com
wrote:

It's simple on a motorway. But, for a simple example on a surface
road, how
do you know where the straight lanes go if there are 3 lanes after an
intersection but 2 before?


Depends on what data you have. If you have lane width information,
then it's easy to figure this out. If you don't, then you can't
figure this out.


Nevermind. Yes you can. If the center of the 3 lane road is to the
left of the center of the 2 lane road, then the lane was added on the
left. If the center of the 3 lane road is to the right of the center
of the 2 lane road, then the lane was added on the right.


Ways are not mapped this way. And even if they were, this would only
work if there's a median.


Not to mention that the line between the lanes doesn't always go 
straight through the intersection.


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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 1:38 PM, Nathan Edgars II nerou...@gmail.com wrote:
 This can be determined by the geometry of the ways, which are mapped
 at the center.

 No they're not.

On Sun, Sep 18, 2011 at 1:39 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 1:38 PM, Anthony wrote:
 Nevermind.  Yes you can.  If the center of the 3 lane road is to the
 left of the center of the 2 lane road, then the lane was added on the
 left.  If the center of the 3 lane road is to the right of the center
 of the 2 lane road, then the lane was added on the right.

 Ways are not mapped this way.

Ways aren't mapped at the center?  Where are they mapped?

 And even if they were, this would only work if there's a median.

What's a median got to do with it?

On Sun, Sep 18, 2011 at 1:40 PM, Nathan Edgars II nerou...@gmail.com wrote:
 Not to mention that the line between the lanes doesn't always go straight
 through the intersection.

Why does that matter?

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 1:52 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 1:47 PM, Anthony wrote:
 Ways aren't mapped at the center?  Where are they mapped?

 Somewhere between the two edge lines. Always using the exact center would
 require zigzagging whenever lanes are created or destroyed.

And mapping wherever and however you feel like it makes for less useful maps.

 And even if they were, this would only work if there's a median.

 What's a median got to do with it?

 When there's no median, the center depends on both directions.

If you don't know how many lanes are in each direction, then this is a
problem.  But you can state how many lanes are in each direction
without mapping each lane as a separate way.

 On Sun, Sep 18, 2011 at 1:40 PM, Nathan Edgars IInerou...@gmail.com
  wrote:

 Not to mention that the line between the lanes doesn't always go straight
 through the intersection.

 Why does that matter?

 Because if the pre-intersection right lane is directly behind the
 post-intersection center lane, but an angled dashed line forces you into the
 post-intersection right lane, this cannot be determined without mapping this
 somehow.

The way is supposed to follow the angled dashed lines.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 2:01 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 2:00 PM, Anthony wrote:
 The way is supposed to follow the angled dashed lines.

 I thought the way was supposed to be in the center between the edge lines.
 Which is it?

I never mentioned edge lines.  There generally aren't any edge lines
in an intersection, are there?

The way is supposed to be in the center of the lanes.  So if lanes=2,
it would be on the dashed line.  If lanes=3, it would be between the
two dashed lines.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 2:07 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 2:01 PM, Nathan Edgars IInerou...@gmail.com  wrote:

On 9/18/2011 2:00 PM, Anthony wrote:

The way is supposed to follow the angled dashed lines.


I thought the way was supposed to be in the center between the edge lines.
Which is it?


I never mentioned edge lines.  There generally aren't any edge lines
in an intersection, are there?

The way is supposed to be in the center of the lanes.  So if lanes=2,
it would be on the dashed line.  If lanes=3, it would be between the
two dashed lines.


And if lanes suddenly jumps from 2 to 3, then what?

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Nathan Edgars II

On 9/18/2011 2:15 PM, Anthony wrote:

On Sun, Sep 18, 2011 at 2:10 PM, Nathan Edgars IInerou...@gmail.com  wrote:

And if lanes suddenly jumps from 2 to 3, then what?


Then you split the way.


That doesn't answer the question. Where do you put the node connecting 
the two ways? At the center of the 2 lane part or the 3 lane part?


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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 2:23 PM, Anthony o...@inbox.org wrote:
 I'm fine with that if we can be consistent.  But that means
 http://g.co/maps/4j2uh is tagged as lanes=0, lanes:turning:left=2,
 lanes:turning:right=2.

Otherwise, lanes=4, lanes:turning:left=2, lanes:turning:right=2 is ambiguous.

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


Re: [Tagging] lanes tag dispute

2011-09-18 Thread Anthony
On Sun, Sep 18, 2011 at 2:21 PM, Nathan Edgars II nerou...@gmail.com wrote:
 On 9/18/2011 2:15 PM, Anthony wrote:

 On Sun, Sep 18, 2011 at 2:10 PM, Nathan Edgars IInerou...@gmail.com
  wrote:

 And if lanes suddenly jumps from 2 to 3, then what?

 Then you split the way.

 That doesn't answer the question. Where do you put the node connecting the
 two ways? At the center of the 2 lane part or the 3 lane part?

There would be one at each.

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