Proposal talk:Winter roads

---

Don't write anything below, type your opinions above. Below is the old discussion, related to older version of proposal, that suggested to use surface tag.

---

Old voting

• I approve this proposal. Glebius 11:13, 6 December 2010 (UTC)
• I approve this proposal. -- Sev 11:34, 6 December 2010 (UTC)
• I approve this proposal. Komяpa 11:36, 6 December 2010 (UTC)
• I approve this proposal. Aleksandr Dezhin 11:37, 6 December 2010 (UTC)
• I oppose this proposal. This is no surface, but attribute of road (winter_road=yes, or winter_only=yes — like bridge or ford) or, less likely, highway type (highway=winter_road). Also, I propose to include in fixed proposal changing surface=ice_road to ice_road=yes --Zversky 15:16, 7 December 2010 (UTC)
• I approve this proposal. Miroff 12:02, 6 December 2010 (UTC)
• I approve this proposal. -- Surly 12:06, 6 December 2010 (UTC)
• I approve this proposal. --Kuprienko Viktor 12:47, 6 December 2010 (UTC)
• I approve this proposal. --Ilis 13:19, 6 December 2010 (UTC)
• I oppose this proposal. Wrong namespace, wrong criterium: The crit is not "winter", the criterium is "below 0°C". So let's say it is highway=whatever with access:maxtemp=0°C --Lulu-Ann 13:25, 6 December 2010 (UTC)
• I approve this proposal. ValentinAK 16:52, 6 December 2010 (UTC)
• I approve this proposal. Dmitry Terentiev 22:23, 6 December 2010 (UTC)
• I oppose this proposal. The surface material is not a "winter road", so it's the wrong key. Propose some other tag. Existence of ice_road (used on 61 ways only) is no reason to introduce more mistaken values. Alv 07:23, 7 December 2010 (UTC)
• I oppose this proposal. wrong use of the surface=* tag, if needed, a tag more in the line of ice_road=yes on a highway or highway=ice_road is more appropriate. --Skippern 10:50, 7 December 2010 (UTC)
• I oppose this proposal. highway=winter_road with surface=ice would be fine. --FK270673 11:17, 7 December 2010 (UTC)
• I say re-think this proposal. Since the surface is changing with the season (ice/mud) usage of surface=* is inappropriate. It doesn't have to have ice it just may be permafrost so one could also name it highway=permafrost_road --katpatuka 12:25, 7 December 2010 (UTC)
• I approve this proposal. Loir 13:50, 7 December 2010 (UTC)
• I approve this proposal. -- Pankdm 13:53, 7 December 2010 (UTC)
• I approve this proposal. -- garl 13:27, 7 December 2010

Not surface

If there's no road most of the year (even if the winter is longer than summer...), just as there isn't anything in the summer in places where skiing pistes run over a lake in the winter, there is no highway. A highway=path, at most, with proper surface=swamp/mud [{tag|wheelchair|no}} etc. Besides describing the physical thing there's at the moment two serious alternatives:

1. go along the skiing piste routes way of tagging. Like a ferry, the road in winter is a nonphysical route otherwise. For example route=winter_road.
2. Or there's a proposal, and some actual tagging already of ways not open in the winter, but in a way capable of describing "open to motorcars in the winter": for example motor_vehicle:seasonal:winter=yes. Alv 13:42, 9 October 2010 (BST)

highway=path is wrong - that's for pedestrian-sized paths, not car-sized paths. Stevage 22:30, 10 October 2010 (BST)

Plain path, without any other access tags, tells us nothing other than "something for navigation, open to nonmotorized" - it says nothing about the width or any other attributes. Without any access mode tags at all, one can only guess that it's more likely something small, since it would be some else highway type if it were a proper road/track/cycleway or whatever. You're right that depending on the location (2nd vs. 3rd photo), highway=track can be suitable, too, as it would be for the 2nd photo - but not for the grass/swamp only 3rd photo. Alv 07:06, 11 October 2010 (BST)

Yes, we already map as highway=track those roads, that can be driven by 4x4 during summer. Even if they officially are/were winter roads. Example http://osm.org/go/8P2reNiN-- --Glebius 11:41, 7 December 2010 (UTC)

This is not like route=piste, where nothing physical exist during summer, or route=ferry, where nothing physical exist at all. During summer winter_road is a cutline, a track, a rut, a path, etc. Winter road can be seen on low resolution satellite imagery. It is a physical line that should exist on topographical map. --Glebius 11:41, 7 December 2010 (UTC)

As each part of the winter road can be tagged as many other things, than a relation would be the most sensible solution. route=* isn't only for bus routes and ski slopes, but can also be used for highway networks (E-roads in Europe as an example). I don't see why a winter road consisting of various cutlines couldn't be tagged as man_made=cutline with a relation route=winter_road - just an example --Skippern 12:52, 7 December 2010 (UTC)

What would be the benefits of such tagging? What tag should be used when winter road exits wooded area and runs upon tundra or swamp? --Glebius 13:37, 7 December 2010 (UTC)

With relation you can put the winter road on any linear member, so sections of river, paths, tracks, cutlines, existing roads, and even untagged lines can be combined into your winter road. This way it can have your special rendering, allow routing and all of this, without interupting the common meanings of the items making part of the winter road. The relation can also hold information about how to find out if the winter road is open (telephone number to local information office), special references, etc. If you want to learn more about relations, see Relation --Skippern 18:20, 7 December 2010 (UTC)

First, I wonder, how and which common meanings do I interrupt when I mark a winter road as winter road? Second, why don't you draw general highways like you suggest: here it goes through wood, mark it as cutline, here it crosses field, let's mark it as embankment, crosses river - it is bridge, enters tunnel - it is tunnel. Let's remove the highway tag from all these objects and create a relation containing them? This can be carryied to the point of absurdity: draw everything with untagged ways and then put relations upon. Finally, most winter roads are drawn using GPS traces. Satellite imagery in far northern regions usually has poor resolution. Mapper have no idea where should he put a man_made=cutline, a waterway=river or untagged line, how you do suggest. Even a man, who had recorded the trace can't tell you what kind of soil (wetland or rocks) was there at a particular kilometer of a winter road. Heh, you can try to disassemble this one http://www.openstreetmap.org/browse/way/72554116 into pieces as you suggest, to see how difficult and pointless that is. --Glebius 19:24, 7 December 2010 (UTC)

I already do mark highways, and generally major highways in relations. This is to tie several pieces of highway into one object, and than make it parts of a network. One example is the Brazilian highway BR-101, which is more than 4000km long. It is drawn as several houndred objects linked together with a relation. Now in a major highway, these object are all highway=*, while your winter road can be recognized as a highway through a relation, as it is loads of other things in the summer time. Anyhow you choose to solve it, as some form of highway=*, a special winter_road=*, usage=*, whatever, surface=* is clearly wrong. I suggested relations because I see that as one possible solution for your special need. It might not be the right one, but for what it is worth, it is clearly more correct than surface=* --Skippern 10:40, 8 December 2010 (UTC)

Okay, we manage route relations to gather highways, but we do not clear the highway= tag from pieces. Because being a highway is the key feature of these ways, although they are also cutlines/embankments/tunnels/bridges/etc. For a winter road the key feature is highway, too.

And you did not reply on the first and third question. Let me repeat them: 1) which common meanings do I interrupt when I mark a winter road as winter road? If I point a northern dweller at a winter road and ask him "What is this?", he will reply me "This is winter road". Not cutline, not track, not untagged line... 3) Please look at examples of real winter roads, and try to tear them into pieces as you suggest, to see how difficult and pointless that is. --Glebius 11:05, 8 December 2010 (UTC)

Correct me if I'm wrong, your suggestion was surface=winter_road, why are you now talking about highway=winter_road? --Skippern 11:35, 8 December 2010 (UTC)

Nope. Suggestion is highway=* (in most cases its tertiary), surface=winter_road. I am okay with winter_road=yes. However, we already have surface=ice_road as approved feature, and on land winter roads should be tagged like them. --Glebius 12:24, 8 December 2010 (UTC)

If you read documentation of surface=ice_road you'll probably see that it is not an approved feature but a rather disputet feature that have some tagging. And how does your winter_road differ from those where the ice_road was meant for? Besides one wrong doesn't make a second wrong is accepted. ice_road should be taken out of surface=* --Skippern 13:38, 8 December 2010 (UTC)

I thought that if ice_road is already in the Surface page, then it is approved. Since it is not, I will transform my proposal to winter_road=yes, ice_road=yes, and again run it through Draft/RFC/Voting. --Glebius 19:55, 9 December 2010 (UTC)

Also not surface

I saw a documentary on these roads (Discovery Channel maybe), so I get what you're talking about. Surface is definitely the wrong tag, because the surface could vary - parts of the winter road could be dirt, parts could be grass, parts could be water. I suggest one of:

• highway=track,track=winter_road
• highway=winter_road

Possibly combined with some other tags like 4wd_only or whatever we have. Stevage 22:29, 10 October 2010 (BST)

Winter road is not track, because it leads somewhere (not of agricultural use). Usually some distant villages have access to the rest of the world only in winter by those roads. --Zversky 12:05, 6 December 2010 (UTC)

Well, most opposals seem to come from people who don't understand what winter roads are. This is okay, since they exist only in Russia and in Canada. Stevage, probably the documentary on Discovery was about some other kind of roads. Winter roads have thick tamped snow as surface always, occasionally bare ice on water crossings. One may suggest "surface=tamped_snow", but I am strongly against it. Let me explain: winter road is a special term, a notion, that describe this kind of road. The term imply surface as well, but describes the road more broadly than just "a road with surface of tamped snow" .In russian we have even special separate word for it - "зимник", with grammatical root "зима" (winter).

Track won't fit here at all, as said by Zversky above.

The highway=winter_road would be inline with highway=motorway and highway=residential, since these words are strong terms that describe the road. I liked that. However OpenStreetMap had came to a conclusion that highway tag represents the importance of the highway, not its physical characteristics. Russian OSM team also accepted that proposal. So, physical characteristics of a road move to other tags, and surface= seem suitable. BTW, most of winter roads in Russia are highway=tertiary. --Glebius 11:24, 7 December 2010 (UTC)

Actually, I agree that winter road should not be surface, but highway=winter_road, since this attribute completely changes routing. Highway class does not matter when road is absolutely impassable. --Zversky 12:36, 7 December 2010 (UTC)

And absolutely passable during other part of year. --Glebius 12:44, 7 December 2010 (UTC)

I've come to oppose this proposal actually. Alv, Skippern and others are right, this is no surface, but attribute of road (winter_road=yes, or winter_only=yes — like bridge or ford) or, less likely, highway type (highway=winter_road). --Zversky 15:23, 7 December 2010 (UTC)

In this case you need to move surface=ice_road away from the approved features, as well, and rename it to ice_road=yes. These two types of road are similar. In Russian language they are even named with the same word. I'm okay with winter_road=yes. The only question where were all of you when the proposal was in Draft and RFC state? --Glebius 15:23, 7 December 2010 (UTC)

Wrong namespace, wrong criterium

The crit is not "winter", the criterium is "below 0°C". So let's say it is highway=whatever with access:maxtemp=0°C Well, I'd usually prefer SI units, but Kelvin is not cool here... --Lulu-Ann 10:15, 12 October 2010 (BST)

Not true. The criterium for the road to be opened or not, is not only todays temperature, and 0°C is not a magic point. Usually, to start the season, winter road requires a few weeks with temperature below -20°C°C, so that all water crossings get frozen to the deep. It also requires large amount of snowfall, so that there would be enough snow to tamp and level. In spring winter roads may function few weeks after air temperature gets above 0°C. The access to most winter roads is driven by authorities, not thermometer. They may close the road temporarily in the middle of the winter, if snow blizzard is expected. --Glebius 10:54, 7 December 2010 (UTC)

Let's go back to the usecase: An onboard navigation computer needs to know, wether to use the road for routing or not. In both cases there is a user dialog needed: mine is for the temperature (I don't stick to 0°, we can also say -20°), yours is for the calender ("Winter"). Both does not work, because the former temperatures are needed, the user must make a decision if the road can be used. With "maxtemp" we have a tag that goes in line with maxheight, maxweight and similar. With "Winter" you don't even go in line with the faily fitting opening hours... What does fit better? I still prefer maxtemp. --Lulu-Ann 13:09, 7 December 2010 (UTC)

Mine is not calendar. Before driving one should determine whether road is open: make a phone call, ask someone living there, ask road police, etc. Onboard navigator computers should not build routes via winter roads by default. --Glebius 13:32, 7 December 2010 (UTC)

Yes, Onboard navigator computers should not build routes via winter roads by default. They should not drop routes via winter roads by default, either. That is what I am saying: We need a dialog to ask the user. Dialogs for maxheight, maxweight etc. exist. They do not exist for the new tag. --Lulu-Ann 13:52, 7 December 2010 (UTC)

Wow, in which navigation software do exist dialogs for maxheight, maxweight etc??? That's a cool feature. I've seen only "Exclude ferries? yes/no". --Glebius 14:16, 7 December 2010 (UTC)

In a perfect world the status would be available over TMC, or a similar future system. Alv 14:42, 7 December 2010 (UTC)

True, but this is also relevant on printed maps, so we need a tag. --Lulu-Ann 12:39, 20 December 2010 (UTC)