Talk:OSM tags for routing/Access restrictions

From OpenStreetMap Wiki
Jump to navigation Jump to search

Missing types of highways

highway=track, highway=service and highway=byway are missing in the tables. The problem with tracks and service roads is that the defaults can't be described with the values "yes" and "no". The following describes the situation in Germany - I don't know the details for other countries: Although many service roads have no formal access restrictions, they could be considered to have access=destination for the purpose of routing - most service roads are simply of no use for other traffic. Motorized traffic on tracks is usually restricted to agricultural traffic. Byways don't exist in Germany ;-) Markobr 07:33, 14 April 2009 (UTC)

We need default values for accessing that streets/ways, becouse mostly the mapper won't make it explicit. --Langläufer 07:25, 16 September 2010 (BST)
I don't see any problem with describing the default access-tag with "yes" or "no" in Germany. In my opinion, there is no reason to
restrict the access on service ways by default for any kind of traffic, so I added a row with all values set to yes.
Towok79 03:36, 12 November 2015 (UTC)
So we'll just add highway=service and highway=track as separate row without value and let the users from the countries get it right.--Jojo4u (talk) 08:18, 11 February 2015 (UTC)

vehicle and motor_vehicle

How should we implement these ew values? --Cbm 14:50, 10 March 2009 (UTC)

Explicit vs. implicit values

I modified the main table differentiating between explicit and implicit values. A cycleway, for instance, is access=no, bicycle=designated. This implies motorcar=no, but this value is implicit (inherited from access=no). This is useful for combining default values with access values of a specific way: if a specific cycleway is tagged as access=destination, this means that motorcar=destination, and not motorcar=no as the previous table suggested. --FedericoCozzi 13:00, 20 February 2009 (UTC)

Germany mopeds on bicycle-ways

Mopeds on bicycle-ways are not allowed in Germany at all. There are no exceptions for "rural areas" in the StVO.

thats not correct. StVO §2.4 ("Außerhalb geschlossener Ortschaften dürfen Mofas Radwege benutzen"). Cheers :) --Cbm 13:00, 14 August 2008 (UTC)
Rural means "ländlich", so I wouldn't translate "Außerhalb geschlossene Ortschaften" with "rural areas". IMHO the correct translation
is "extra urban areas". --towok79 1:20, 19 October 2015 (UTC)
I am not sure if "extra urban areas" is a common wording in english (I am from Germany). I would prefer the translation "outside of urban areas" instead of "extra urban areas" (as a german user), because the word "extra" is used very much in the meaning of "special" or "especially"/"particular" in Germany, and that would be the exact opposite of what it should mean, like "especially in urban areas". I guess many german readers could understand it like that or could be confused at least. "outside of urban areas" would make it clearer, if this is also OK for native english speakers. Could this be changed? --Goodidea (talk) 15:41, 27 July 2020 (UTC)

Does it make sense

I don't think you should programm rules for different states. What we need is an information in the map, which rules apply for a country area. This throws up the following problem: How does the routing software know about a general access rule or speedlimit for a country, if no part of the country outline object is downloaded in the tile you are using. --Lulu-Ann 15:57, 14 August 2008 (UTC)

How is a routing application going to do routing using only tiles? So routing applications are using raw OSM data and country boundaries are/should be in the OSM data. --Lambertus 12:11, 4 October 2008 (UTC)
What's meant is (if I guess correctly), that if the software or platform doesn't have local storage but downloads only the data for an area between the endpoints, country boundaries are even likely not to be in there. I wouldn't even try to implement such a software as even for a narrow slice the amount of data grows too fast even when it were limited to ways with highway=*. Alv 12:33, 4 October 2008 (UTC)
I think that different access rules for each country is the only way forward (or you'll end up with huge discussion for whether highway type X should allow Y etc, and basically, country traffic rules are too different to combine them in one size fits all shape). I don't think it's a difficult problem to let a router know in which country a certain location is. --Eimai 13:42, 4 October 2008 (UTC)
I just had this experience, my gps with routing sent me over a trunk. It was in bicycle mode and I was in germany, FAIL. With different access restirictions per country this will always happen. --Meier 08:53, 4 May 2009 (UTC)

Pedestrians on cycleways vs bicycles on roads

One question that pops up for me about this page... I made the notes for Belgium that a pedestrian can walk along a cycleway if there is no parallel path for them (so a value of "foot=yes" can be assumed for cycleways). But the same argument counts for roads: you can walk on the road if there's no path where you can walk, and you can cycle on the road if there is no cycleway. So shouldn't the values for roads for pedestrians, cyclists, mopeds and horses be of the orange kind if the pedestrians on cycleway is tagged orange? --Eimai 17:25, 14 August 2008 (UTC)

IMHO is 'yes' correct. because you're not forced by law not to walk on existing roads (forbidden by law). as you said, it's allowed. I don't know the belgium situation in detail (I'm assured you get it better :) ), but e.g. in Germany as a pedestrian you can even walk on a road if you like. There is no law that forbids it.
you catch the situation (orange fields) very good, I think. maybe no.3 is rather a 'no' because the question is: are bicycles allowed even if there is no sign F103? If not -> Conclusion: without allowing bicycles (sign F103) bicycles are not allowed in pedestrian precincts. I hope you can follow my arguments --Cbm 01:34, 15 August 2008 (UTC)


a cycleway is a way designated for bicycles. e.G. in Germany designated cycleways are only for bicycles excempt from it's allowed by additional signs (in OSM normally tagged as foot=designated or foot=yes)
Yeah, but that's Germany. In Belgium these signs with pedestrians on it were only introduced some years back, so the rules still say that a cycleway can be used for pedestrians if there is no parallel footpath. And that rule is exactly the same as cyclists and pedestrians that aren't allowed to use the road itself when there's a footpath or a cycleway on that road. So I wonder where the difference is that we can use "yes" with the roads with no strings attached, but not with cycleways. --Eimai 18:37, 14 August 2008 (UTC)
The Netherlands are mostly the same as Belgium, except that there is no single sign that shows both a bicycle and a pedestrian. In general you can walk on a cycle path if the is no footpath next to it.
"In Poland it's forbidden unless there is no pavement or roadside.", private correspondence --Phicoh 11:00, 13 October 2008 (UTC)
"In the UK, pedestrians can use all highways except where explicitly disallowed", David Earl in osm-talk. --Phicoh 10:01, 15 October 2008 (UTC)
"1/3 of all highway=cycleways have foot=yes in Sweden, I can't remember any cycleway that dissallowed foot access,", Erik Johansson in osm-talk. --Phicoh 11:06, 15 October 2008 (UTC)
is there any evidence by law or is this only an unproved quotation? --Cbm 12:45, 15 October 2008 (UTC)

Belgium

  • There's no reason for a "designated" access tag in Belgium as there is no reason why one has more rights over the other on any of these highway types when different vehicle types have access to a road. "designated" is therefore synonym with "yes"
But using bridleway/cycleway/footway automatically implies a kind of designation.
Otherwise all these ways are only highway=path with some access=yes-Tags.
highway=footway -- synomyn --> highway=path + foot=designated
highway=cycleway -- synomyn --> highway=path + bicycle=designated
highway=bridleway -- synomyn --> highway=path + horse=designated
Well, that's then the world default for these footway/cycleway/bridleway shortcuts. In Belgium they mean something else then. So you could say highway=cycleway = highway=path + bicycle=yes + foot=yes. Nothing wrong with that, is it? There's just no reason why a designated tag would bring any additional information about roads in Belgium ("you have access or you don't") so we discard it. --Eimai 14:59, 17 October 2008 (UTC)
Designation has nothing to do with priority of traffic-members.If there wouldn't be any designated-character in cycleway/footway/bridleway the accuracy of discrimination with signs like Zeichen 237 - Sonderweg Radfahrer, StVO 1992.svg and Zeichen 239 - Sonderweg Fußgänger, StVO 1992.svg would be adsurd.--Cbm 06:11, 20 October 2008 (UTC)
What is designated about then? If this isn't telling anything about access rights on the road (priority or right of way) then what is it doing in the access tag at all?
Those signs just tell what's allowed where and what isn't. The cycleway sign allows mopeds class A as well here, so if bicycles would be designated, why wouldn't mopeds class A be? If a road or path allows a vehicle type, it's by definition designated. The road may not be well designed for that particular usage, but that's another issue, not relevant to the access tag. --Eimai 12:08, 20 October 2008 (UTC)
In the first step a sign like Zeichen 237 - Sonderweg Radfahrer, StVO 1992.svg just shows us: "This way is made for bicycles". It's deSIGNated for a specific type of traffic-member (bicycle=designated).
In the second step it may become special local or country-specific access-allowance; e.g. in your special case Belgium: "Mopeds are allowed to use ways which are designated for bicycles" (moped=yes)
access=designated not only tolds use about the ability of using a "highway", but it's also a quality-attritute for with main-usage this "highway" is made for.
maybe Tag:access=designated helps for a better understanding --Cbm 17:04, 20 October 2008 (UTC)
No, the meaning of Belgian road sign D07.svg is "cyclists and mopeds class A have to use this part of the road (and depending on situation, it allows/obligates drivers of mopeds class B), and other vehicles cannot". So they're for cycle lanes belonging to a road. The traffic rules don't say anything different for mopeds class A and cyclists in relation to that sign, yet you say that it's bicycle=designated but not moped=designated because the sign happens to show a bicycle and not a moped?
Isn't a Belgian road sign D07.svg signed ways/lane made for bicycles? If yes the way is "designated" for bicycles ;)--Cbm 00:43, 14 November 2008 (UTC)
I just don't understand what extra meaning you try to give access=yes with designated. If a road is designed for certain vehicle types that's not an access property, that's a road property. --Eimai 18:05, 20 October 2008 (UTC)
If in Belgium this sign designates a path for bicycles and mopeds as well then tag the path both with bicycle=designated and moped=designated. I didn't say it is designated only for one vehicle-type, but at last it is designated at least for one group of traffic-members. :) (e.g. highway=motorway is designated for any "strong-motorized" verhicles; means motorcars, cars, hgv,...) --Cbm 19:45, 20 October 2008 (UTC)

Finnland

In Finland most ways are of the combined type Zeichen 240 - Gemeinsamer Fuß- und Radweg, StVO 1992.svg and the signed cycleways Zeichen 237 - Sonderweg Radfahrer, StVO 1992.svg are rare, too (re Sweden), but the wording in the definition for that sign "Muut ajoneuvot eivät saa käyttää ..." ("Other vehicles may not use") doesn't forbid pedestrians. If there's a signed footway Zeichen 239 - Sonderweg Fußgänger, StVO 1992.svg in the direction one wants to walk, the statute is "jota jalankulkijan on käytettävä" ("which a pedestrian is to use") - probably intentionally defined so vaguely that it doesn't really forbid anyone from walking on a cycleway. There's even, elsewhere in the law, a statement that a pedestrian on a cycleway "should usually use the side of the cycleway". Banning walking definitively would require a Zeichen 259 - Verbot für Fußgänger, StVO 1992.svg. Alv 14:38, 15 October 2008 (UTC)
so what we defintly can say for Finnland (looks like it's the same in Germany):
Zeichen 240 - Gemeinsamer Fuß- und Radweg, StVO 1992.svg is highway=path bicycle=designated foot=designated
Zeichen 237 - Sonderweg Radfahrer, StVO 1992.svg is highway=path bicycle=designated (Shortcut => highway=cycleway)
I don't see how that follows. The above text suggest that the cyclepath can be used by pedestrians (unless there is a nearby footpath). So, from a routing point of view, assuming 'foot=yes' is a much better default. --Phicoh 18:06, 15 October 2008 (UTC)
Zeichen 239 - Sonderweg Fußgänger, StVO 1992.svg is highway=path foot=designated (Shortcut =>highway=footway)
--Cbm 15:13, 15 October 2008 (UTC)


I wanna catch some thoughts about the finish-access table.
->(1): why it is in the path-row and not in as a note in the cycleway/footway row?
->(2-4): any vehicle_type and have additional allowance (signs) to enter any type of road. but this should not be catched in this table. This table should catch the situation if there isn't any sign. under this premiss, I think, (2) and (4) should be no. the (3) filed should just be orange because it isn't clear how to catch this "either or" situation.
what do you wanna catch with "destination" and "private" for the finnish situation?
looking forward to you explanations :) --Cbm 16:34, 16 October 2008 (UTC)
(1) A horse rider is compared to a vehicle and is not allowed to use a designated footway or cycleway (but those can be, but very seldom are, designated for horses equally). If it's an unsigned path, horses are just as allowed as pedestrians. Alv 17:33, 16 October 2008 (UTC)
(2) I have removed the note, it was unnecessary. Mopeds used to be allowed unless forbidden, but that was changed like 15 years ago when most of the cycleways had an extra "no mopeds"...
(3) When in some cases a pedestrian is told by the law to use the cycleway, it would seem pointless not to say that they may do so, even without explicit tagging which might be forgotten or lost.
(4) If the road (highway=pedestrian) is signed (only) with our equivalent of [Image:120px-Zeichen_242.svg.png|40px] (which was only introduced in 2006), motorcar traffic is explicitly allowed if it's the only route to premises but not for traffic solely onto the pedestrian road. In other words, you may drive through to your destination, which could be some inner yard on a block next to the pedestrian road, but you may not drive there to unload, wait, turn around or whatever. All pedestrian roads were and some might still be signed with a [Image:120px-Zeichen_260.svg.png|40px], which did prohibit motorcars altogether (but usually had extra signs to allow service traffic). Some plazas now tagged as a pedestrian road + area=yes are really just an extension of the sidewalk and those just don't allow motorcars. Without reviewing all ways tagged as highway=pedestrian no one could say what is the correct value for motorcar access, for now.
As to the "private" on path, there's a separate law about motor traffic outside roads which basically requires land owner's permission (owner can be a person, municipality or the state). The land owner is allowed to drive a bus around the hiking trails he owns, even when not feasible. And if it were something big enough to have a sign "no motorized vehicles", it'd be tagged as a footway or cycleway.
Here a living street sign explicitly allows any motorized traffic "to a destination along that road and for parking" (but parking within a living_street is limited to marked areas). (It actually doesn't say that the intended parking must be along the living_street, but I guess proving that in a court could be a challenge). Alv 17:47, 16 October 2008 (UTC)
isn't it possible to use a living_street as a "shortcut" A -> living_street -> B ? If yes, it's not only access for destination, is it?. --Cbm 18:01, 16 October 2008 (UTC)
It's not allowed here: "vain kadun varrella olevalle kiinteistölle tai pysäköintiä varten" -> "only to premises along the road or for parking". Alv 18:16, 16 October 2008 (UTC)

Restrictions on hgv

In my area hgvs are allowed only on (the equivalent of) "primary" and "secondary" roads, as well as any other roads marked with "Truck route" (I've considered the latter to be equivalent to hgv=designated). Is this common enough around the world that those might be good default access restrictions? A map (with allowed HGV routes in maroon) can be found here: [1]--Hawke 17:15, 8 October 2008 (UTC)

(I'd tag them hgv=designated, too.) I would doubt it, at least outside the very urban areas. In my experience in Europe the restrictions are always posted. The most common limitation I encounter locally is a maxlength=12 (with a text sign "doesn't apply to buses") (effectively disallowing only hgvs), but it's used almost only in the city center and in the suburbs or the countryside the restrictions are very rare. I looked around my local area and might add that here in the city center I could only say that a primary is almost always allowed for hgvs and for a secondary the probability is 50/50. Alv 12:51, 9 October 2008 (UTC)
It is not common enough. Usually all kinds of motor-vehicles are allowed on all kinds of motor-roads. It should thus either be tagged explicitely on the way-level (easiest to use for routes) or a rule applying to the country as a whole (we cannot yet determine the current country due to often incomplete and broken borders).
What is "your area"? Federal State or just a very local district? Because for a global-wide default it's not common enough. --Cbm 12:02, 12 October 2008 (UTC)
"my area" in this case is the city in which I live -- so, very local. Just to confirm, you're saying that in most places, HGVs are allowed to go anywhere that there's no sign forbidding them? --Hawke 19:42, 24 October 2008 (UTC)
yes. E.G. in Germany if there isn't an explicit restriction HGVs, goods and busses can go anywhere a normal motorcar can go, too. --Cbm 21:38, 24 October 2008 (UTC)
Yes. I see HGV going on all kinds of "motorway", "trunk" and even "residential" and they act perfectly legal (Germany, Austria, France, Swizerland, ...). --MarcusWolschon 14:11, 20 February 2009 (UTC)

values other then "yes" and "no"

Please keep in mind that these are truth-tables supposed to be used by computers, not humans. I removed the "designated" as this does not imply any special treatment for a routing-engine. A way is part of the graph or it is not. --MarcusWolschon 14:04, 20 February 2009 (UTC)

I strongly oppose this edit. First of all, this is not a truth-table: this table simply summarizes OSM default access restrictions (which for instance state that bicycle=designated on a cycleway). A computer software can easily translate designated into yes if it so chooses. There is no reason to lose that information on the wiki page itself.
Moreover, only very basic routing algorithms assume that the road network is a connected/unconnected graph. More complex algorithms assign weights to the different ways (e.g. favour motorway over unclassified road) and the difference between designated and yes can be exploited by algorithms (see e.g. A*-algorithm).
I am going to revert your changes.

No, computerised routing algorithms are not the only "end user" for mapping data. Renderers, and the human users that see the rendered map, can still benefit from designated ways which may be shown differently - different colour or line/dash style. Such differentiation can be useful when multiple ways, parallel or close-by, all end up having access=yes (foot=yes, cycle=yes, etc.), perhaps by inadvertence. Although several ways may be possible, one of them could clearly be preferable, and should be "designated" as such. A typical example is a footway/cycleway/bridleway lying beside an ordinary road (instead of or perhaps in addition to a common bike-lane/sidewalk). Sometimes the road might actually require a foot=no tag or etc., but often this is not strictly-speaking justifiable, as the footway/cycleway/bridleway is simply preferable in most cases, not mandatory, so that a user could sometimes have a good reason to follow the main road. --Neil Dewhurst, Lyon France (talk) 17:21, 2 January 2022 (UTC)

trunks (especially in Germany)

In Germany trunks are tagged with motorroad=yes if they are only for motorvehicles. So trunks without motorroad=yes may also be used by bicycles. --chris66 22:23, 27 December 2009 (UTC)

they should be tagged with, but often they don't! --Langläufer 07:22, 16 September 2010 (BST)
Hm?
I really wonder why the trunks are said to be usable by bicycles and pedestrians by default. Where does that information come from? I cannot remember any trunk type road that I would have cycled or even walked on without being catched by the police. The issue is that some navigation engines like Mapzen and probably also Mapfactor use the information from this page and provide walking and cycling routes on trunks. This is - again - wrong and not possible in most cases in Germany I think. If there are walking and/or cycling ways along a trunk they are normally tagged or even drawn explicitly.
--Krza (talk) 20:34, 24 July 2017 (UTC)

path and road

I'm not sure why path and road are in this table.

highway=path is, by default, "non-motorised traffic only". You then modify it using the access tags. If you're going to start having country-specific defaults for highway=path, that defeats the point of the tag and you might as well use highway=cycleway or highway=footway.

highway=road means "there's a road but I don't know what it is". If you don't know what it is, you can't tag what access restrictions there are. If you do know what the access restrictions are, then it shouldn't be highway=road.

So I'd suggest removing these rows from each table.

--Richard 12:09, 15 September 2010 (BST)

Full agreement. RicoZ (talk) 22:19, 3 July 2015 (UTC)
I don't agree. IMHO a highway=path is a narrow way, presumably suited to only foot, cycle, horse (mounted) or moped/motorcycle traffic, perhaps subject to restrictions. Cars, trucks and tractors, etc. are wider, with wheels on left and right, and require highway=track or other highway=yes/unclassified/residential/*, perhaps subject to restrictions. I think its appropriate to map paths as such if and when they are actually seen to be narrow. --Neil Dewhurst, Lyon France (talk) 17:21, 2 January 2022 (UTC)

formating of tables

i think, the formatting of the tables should be changed. row and columns should change places, since highway values stay the same, but access values keep on growing. --Flaimo 11:37, 1 May 2011 (BST)

also i think it would be better to put the table of each country on its own wiki page. users could then discuss the default values for each country separately, while this page here links to the individual pages --Flaimo 15:06, 1 May 2011 (BST)


Is that correct, or spam

In Russia section references seem to untrusted web addresses. The descriptions are Russian and the references section has no heading. This should be checked. -- Gisbert (Gmbo) (talk) 06:02, 25 March 2015 (UTC)

"permissive" in the table?

This might be useful in some places for paths or tracks, any problems using the value in the table?

Is there a list of tools using this page, automatically parsed or human encoded ?RicoZ (talk) 09:53, 8 June 2015 (UTC)

Default values for *_link

Is it fine to assume that highway=x_link has the same default values as x? For example highway=tertiary_link. I see in the data that often, there are is no explicit access defined for *_link values.

Yes, e.g. the sign with the access restriction of a motorway in Germany (car has to be able to driver faster than 60 km/h) is on the motorway_link. On the other side this might expose more tagging errors. E.g. see http://www.openstreetmap.org/way/170704283/history which prevented crossing the motorway.--Jojo4u (talk) 07:19, 11 June 2015 (UTC)

Construction=yes removed

I removed construction=yes as "under construction" in the default section. This is not recommended and bad mapping practice. It will often be forgotten after completion and imho it's not recommended to obey it for a router. There are only 3000 construction=yes with highway=* compared to 85000 highway=construction.

Problems

This page has some problems:

fundamental problems

  • several entries contradict the approved and valid OSM definitions of the respective features
  • the table is not fine grained enough:
    • default properties of highway=track and highway=path will depend on the value of informal=*
    • default properties of highway=track and highway=path depend on land ownership and landuse, in general it is impossible to give a nationwide default
  • some concept of traffic zones may be needed
  • the table does not know the value permissive

practical problems:

  • many entries indicate that contributors confuse the meaning of "default access restriction" (meaning of the "if not otherwise posted")

--RicoZ 13:32, 13 August 2015 (UTC)

Please could you be more specific:
Which entries contradict which definitions?
Where do you see a dependence to informal=*? That key is neither mentioned on the page, nor approved, nor in use.
Why do you consider it a problem that the page does not know the value "permissive"?
ownership and landuse: Just choose a default that matches most cases. That's what a default is meant for. A default does not need to cover all cases.
--Fkv (talk) 12:47, 13 August 2015 (UTC)

Attention

This proposal here creates big confusion by mixing two completely independent questions: (1) what is the legal access if an OSM way has no access tags, (2) what is the legal access if a (real world) way is not signposted. While there are still some mappers insisting that both should be similar, since many years mapping praxis established OSM default rules (independent of national regulations). So please read the national tables on this page with extrem caution and with the fact in mind that it's not clear what question they answer. And if you are missing the default access for "track" and "service" in the top ("default") table in this page, please do not look in "your" national table, but look in the map for examples. --Arndt (talk) 11:50, 26 May 2019 (UTC)

Please explain what you mean by "it's not clear what question they answer". The defaults are given as documentation for application developers. You don't need to wait for a question before you can document something.
Please also eleborate on why you suggest to ignore country-specific defaults for track and service. When there is a default given, you can expect that there is a good reason for it. --Fkv (talk) 22:52, 26 May 2019 (UTC)
isn't that what my comment was all about? Or read the section "Does it make sense" earlier on this page here. I also talked abaut the issue at this years FOSSGIS conference, maybe you wan't to have a look at the video (german): https://youtu.be/eO-9wW1aOJI?t=381
The last comment in the "Does it make sense" section was by Meier who stated that "different access rules for each country is the only way forward". For more than 10 years, nobody has ever objected. Now you base your rejection of country-specific defaults upon a discussion that concluded that these defaults are necessary?
No, I won't watch a 21 minutes long video. Please make your point here. A few sentences should be enough to explain what you want and why. --Fkv (talk) 09:49, 27 May 2019 (UTC)
I want a routable map and that requires coherence between mapping praxis and wiki theory. My (few-sentences) initial comment here is the compact english veriion of the 6-minutes of my Dresden talk starting from the start-pointer (6:21) I posted here
Where does mapping practice differ from wiki theory? Please be more specific. In general, mapping practice should conform to the wiki and vice versa. --Fkv (talk) 11:48, 27 May 2019 (UTC)
For a few years the article has started with "if it becomes a generally accepted practice." I don't see this happening currently as the main routing systems using data from OSM (OSRM, GraphHopper and OTP) do not support rules by area and don't seem to have that feature in their roadmap. Mappers would prefer this as it saves some work while creating ambiguity (why are access tags missing, were they not surveyed or is the mapper uncertain? should surveyors re-survey the ways to make sure?). I think this article should be used as an incentive towards standardization of mapping practices between countries, to reduce the need for country-specific rules.--Fernando Trebien (talk) 13:46, 27 May 2019 (UTC)
The "generally accepted practice" refers to tagging practice, not to what applications do with the data. We as mappers provide all the necessary data and documentation for applications to implement country-specific rules. What priority application providers assigns to the implementation of these rules is up to them to decide. I guess that one of the reason why they deprioritize it is that their users don't nag them with it. Users of the routing engines on the OSM website report routing errors as map notes, because they are not aware that routing engines have their own bug tracking systems. There is no reference to them on the OSM website.
As to ambiguity: This is an issue with mapping in general. When a house is not mapped, data users cannot be 100% sure that there is no house. It's possible that there is a house, but no mapper has ever come across it. It's also possible that there are mappers around but they haven't mapped it because it does not exist. A third possibility is that no mapper has ever surveyed that place but there's no house anyway. Whenever in doubt, data users should assume that the data correspondends to what can be found on the ground, i.e. when a feature is not mapped it is not there, and when no access tag is set there is no evidence for access permissions or restrictions. In that case, the country-specific laws apply, and that makes country-specific defaults so important. --Fkv (talk) 15:09, 27 May 2019 (UTC)

Rationale for United States Access Restrictions

I'm getting into the weeds of tagging some paths on campus, and I've been referred to the table here for the U.S. tagging conventions. I don't know if this is just me, but why is moped=yes an automatic default for highway=path? It seems a bit counterintuitive as mopeds (at least in my area), most definitely are only allowed in the same roads and streets as cars. I'm using this as a justification to not use highway=path, as in my opinion, the automatic designation value given to each of the other kinds of paths (footway, cycleway and bridleway) address the fact that the paths are almost always designed and built primarily for a kind of end-user, despite their ability to accommodate other forms of traffic. Thoughts?--IanVG (talk) 17:22, 30 September 2020 (UTC)