Proposed features/Passability

From OpenStreetMap Wiki
Jump to: navigation, search
Passability
Status: Abandoned (inactive)
Proposed by: *
Tagging: passing=normal/ caution/ difficult/ impossible
Rendered as: varied thicknesses of how an unclassified road renders


There are 4 widths of road commonly found that effect the speed a car drives at and therefore I think its inportant to tag them so a route planner can take them into consideration. Otherwise if its told just to avoid all unclassified roads It would be poor for people getting around in areas where there are far fewer main roads, and if its told to use them then it would be a pain to find yourself getting taken down the smallest of lanes. A roads width could always be added manually if you got out and measured it, but for the purpose of making it easy to set up a few set rules for rendering, and just being realisitic, cause I dought anyone will really measure raods; I am proposing the following predefined road sizes.

After a converstaion on the mailing list it was suggested the following would work:

  • Passing=normal: For use on a road that usually about 3 cars wide, so provides space to pass. The roads usually have road markingS. Average speed is 50-60mph. You don't slow down to pass the average car on the other side
  • Passing=caution: For use on a road that is about 2.25 cars wide. Usually the road has no markings as there isnt enough room to create 2 spacious lanes. Average speed 40-50 When passing another car you would shift to the side and slow down slightly, but otherwise just keep going. Usually you are forced over suffiecently that you are made to drive in the damanged side areas of the road wich are usually outside the white lines on bigger roads.
  • passing=difficult: For use on roads that are about 1.5 cars wide, but with the area next to it where the surface is ok to drive onto in usual weather, meaning you can go up onto the grass to pass. Usually roads like this have frequent pass places in the UK. Average speed is usually about 25-30mph.
  • passing=imposible: For use on roads that are again about 1.5 or under cars wide, but With high banks, or walls etc either side wich means that 1 of the 2 cars must reverse. The roads are so tight that overtaking horses or bikes is usually inposible without a passing place, or a slowing to standstill. Average speed is about 15-20mph.

I'm not suggesting a road chopping and changing every 100 yards, but rather average it over a mile or so. The definitions of these tags can easily be transfered to be more fitting to different countries and terrains I think.

This tag would be additional to highway=unclassified, so its not additional highway tags, and its not a replacement. Ben. 04:37 21st January 2007 (UTC)

Comments

  • I do not see that this adds anything more to the data than is already provided and can be achieved by the pre-existing 'width' and 'lanes' keys, as a result I shall not be supporting it -- Batchoy 11:58, 24 January 2007 (UTC)
    To quote myself. "A roads width could always be added manually if you got out and measured it, but for the purpose of making it easy to set up a few set rules for rendering, and just being realisitic, cause I dought anyone will really measure raods"...I have proposed this. Lanes is just plain incorrect. Ben. 18:14 24th January 2007 (UTC)
    • However the proposal is predicated on an esoteric measure called 'car width' and as such is open to misinterpretation; the passability for a Smart ForTwo (1.5m wide) is vastly different to the passability of a Hummer (2.2m wide). For the key to be of any value 'car width' must be defined in terms of SI units, say quite arbitrarily 1.8m. Having done this you judge the with of the road to be 1.5 car widths wide you may as well put 2.7m as the road with rather than passing=difficult. Knowing the road width to be 2.7m is of greater use to all road users, than knowing that passing is difficult for road users with vehicles about 1.85m wide. -- Batchoy 09:18, 25 January 2007 (UTC)
  • It's open so that it can be adapted to each county. Car widths vary, so If we were to state that the unit of 1 car width is x feet/meters then that would be pointless if a county has cars that are twice as wide as anywhere else, because suddenly a passable would be an inposible. In the majority of the world cars are similar enough so that a good estimate can be made of how easily the average car can go down a road. If a roads width is known to the cm then a person could just add width, although I think width is less practical for rendering. Usually people could only know a road to about a 50th of that acuracy. Stating the roads width as a measurment defeats the entire point here. The counter arguement may be to have a set of set default widths to use if we don't know the exact road width, but then why add inacurate data? This proposal just groups similar widths, and also takes other things into consideration such as ditches and banks on the sides. Ben. 15:08 25th January 2007 (UTC)
  • I believe there is a role for qualitative measurements in OSM so that we mappers can make recordings without have to stop and whip out tape measures or plumb bobs. A "car width" seems to me a reasonably self-explaining measure, though to be sure I suggest that it be defined as "the width of an average family car in the country mapped". I therefore support both passibility and lanes and let the market decide by rate of actual adoption. lanes to me is, or should be a synonym for car widths. passibility does give information than just lanes. A long thin straight road through a desert can be travelled at 120 kmph+ and passing even large trucks involves merely one vehicle bumping off the side of the road for a few metres. A winding walled road of the same width in Yorkshire, UK will probably involve being stuck behind the smallest car. I also have issues with what the current definition of lanes means (one this way, two that way = 1 or 3?). MikeCollinson 04:20, 26 January 2007 (UTC)

  • The discussion about this goes on and on in the IRC, and so heres where its at.
Maybe a solution would be to redefine the 'Width' key so it is only as exact as the tags used are set to mean. So, we have:
1) Just the road type, just as highway=unclassified. Just blank, the most basic way.
2) Basic tags: Width=double or Width=single get added. So theres thick and thin ways.
3) Moderate accuracy tags: Width=Normal|Caution|Difficult|Inpossible so theres a range of thicknesses.
4) Super accurate tags: Width=5.52(meters) so theres infinite thicknesses.
Double would mean the same as Normal or Caution which would then mean the same as all +4ish meter measurments.
Single would mean the same as Difficult or Inpossible wich would then mean the same as all -4ish meter measurements.
The other concern is that normal|caution|difficult|inpossible might get muddled with how hard a road is to actually navagate, rather than just how hard it is to make steady progress along it. Wide | Average | Thin | Super_Thin or something to similar effect may helpf solve this problem.
I'm not that convinced that the width tag should be used more loosely, but I'm not shore where else it would be more fitting without having a new key. In the end, I'm not that fussed so long as the tagging scheme doesn't get made even messier, I just really want to get on with splitting these 4 types of road appart. Ben 03:14, 12 April 2007 (BST)
The above makes sense, but I'd not bother with numbers above 2 listed there, it's too dynamic and subjective (not 4 obviously). Bruce89 14:43, 12 April 2007 (BST)

In Korea I've seen what I would call single lane roads with no shoulder that have a yellow stripe down the middle. I'm not sure what that means, but it does mean that the number of lanes without a width can be confusing.

I think meters for the width of the road, not including the shoulders, should be fine. Most roads will be of a handful of standard widths, so once you've measured widths of the handful of standard roads in your country you should be able to eyeball it. --Korea 20:55, 7 May 2008 (UTC)