Why have you changed the http://wiki.openstreetmap.org/wiki/Tag:highway%3Dfootway page so that it is now incorrect?
The statement "highway=footway is mainly used for urban paths designated for pedestrians only. " is simply not true. "highway=footway" is used for a variety of paths which are used mainly or exclusively by pedestrians. There's nothing explicitly "urban" about them.
The wiki should follow how people map, not the other way around.
As I stated in the summary, I tried to improve the distinction of footway vs path. As you surely are aware highway keys like footway, bridleway and cycleway are ancient history of OSM. They combine way and access attributes in a single key. Today they are almost obsolete, because we are able to use generic keys like path with specific access restrictions (i.e. highway=path + foot=designated + horse=no + bicycle=yes). As to my impression globally it seems to be correct pretending that designated footpaths exists mainly in urban or residential areas. I wasn't aware if this does not apply to the UK.
The key reason why people use "highway=footway" isn't to "combine way and access attributes", it's to provide a description of what the way in question is actually like. "highway=path + foot=designated + horse=no + bicycle=yes" tells you nothing whatsoever about what the path in question is physically like at all. It says that foot traffic is supposed to use this route rather than some other (undefined) one, and that legally cyclists can also use it but horse-riders cannot; but nothing else. If instead footway, bridleway or cycleway had been used instead it would give a clear indication to users as to what sort of path they should expect. With a bridleway a horse-rider would expect gates rather than stiles and enough clearance to get a horse through; with a cycleway a cyclist would expect e.g. a surface that they can cycle on and gates or barriers that they can negotiate with a bike. Note that this is entirely separate from the legal access rules that may apply. The key descriptive part from the old footway wiki page was "... used mainly or exclusively by pedestrians. ". If you live in a first-world city and never see the wider world then perhaps you think that in all places where a particular mode of access is legally allowed it's actually physically practical, but on most of the planet this simply isn't the case.
Encouraging people to use "highway=path" instead of proper descriptive definitions like footway, bridleway, cycleway etc. encourages them to not capture as much detail as they otherwise would. There may be occasions where none of footway, bridleway or cycleway fit, but they're rare. If people change highway=footway to highway=path (which someone did after seeing your changes to the page) then they're actually deleting useful information from OSM, and that's not good. --SomeoneElse (talk) 23:10, 21 May 2015 (UTC)
Actually - one more thing. I just remembered this diary entry from a year or so ago:
It describes the problem perfectly.
On first sight, descriptive keys like footway or cycleway seem to give a clear indication of the type of a path. Actually that’s only true, if you provide - in addition to the highway tag - other proper attributes, such as surface, width, smoothness, incline etc. So either way an equal number of tags are usually required.
The reason why I changed
„…used mainly or exclusively by pedestrians“ to
„…mainly used for urban/residential paths designated for pedestrians only.“
is, that on most of the planet (outside the UK ;) there are only few sign posted footways out of residential areas. Rural paths may be used by all kinds of non 4-wheel traffic, including stock, mule, yaks, motorcycles etc. and are rarely footways.
The actual usage of footway vs path in the UK is differently from most of the rest of the world, partly because of varying default access restrictions. So I added a pointer
„If you are mapping footways in the UK (specifically England and Wales) see http://wiki.openstreetmap.org/wiki/United_Kingdom_Tagging_Guidelines“ and changed urban to residential.
I appreciate the insight you provided, I was not familiar with UK-specific tagging.
There's nothing "UK specific" about the fact that "highway=footway" (or "highway=cycleway" etc.) provides more information about something than "highway=path". If you haven't surveyed properly (or in rare cases when none of the other options fits better) then "highway=path" may be appropriate, but you can't argue that persuading people to use "highway=path" rather than "highway=footway" removes information from the map. Please don't do this. You say that "is, that on most of the planet (outside the UK ;) there are only few sign posted footways out of residential areas" but I really don't see what "signposted" has to do with it (most of the planet is not like Germany). You say that "Rural paths may be used by all kinds of non 4-wheel traffic, including stock, mule, yaks, motorcycles etc. and are rarely footways". That may be true in some cases; if so don't map that as a highway=footway - use a more appropriate tag. Actually, I've seen motor-vehicle restrictions on what you'd call "paths" on 3 of the 4 continents that I've mapped on so I'm not convinced that it is the general case. If local mappers (such as in Germany) have decided to use "cycleway" as a special case for foot=no cycleways and "path" for foot=yes ones that's parseable by data consumers on a country=by-country basis, but if local mappers elsewhere start thinking that there's something inherantly "residential" about highway=footway (as happened since your edit) then map information is degraded as a result and we're all worse off.
I appreciate your consideration and have changed http://wiki.openstreetmap.org/wiki/Tag:highway%3Dfootway
IMHO it’s not a proven fact that "highway=footway" (or "highway=cycleway" etc.) provides per se more information about something than "highway=path“. You could technically replace all highway=footway/cycleway/bridleway with highway=path and according tags for access and physical attributes. There’s nothing „ill-considered and (…) dangerous“ about recording physical and access tags separately and unambiguously. Actually it’s essential to handle the plethora of different national access-restrictions in a global database.
At least for the more common standard keys tagging-presets of the main editors will affect how the tagging goes. My edit to the wiki is consistent with the the current presets in iD and JOSM. I hope my edit is a reasonable compromise. --geow (talk) 21:12, 28 June 2015 (UTC)
I've no idea what "designated for pedestrians only" is supposed to mean. Does it mean "if there's no sign it can't be a footway"? The original wording was "mainly or exclusively by pedestrians", and that is far better. --SomeoneElse (talk) 14:07, 15 July 2015 (UTC)
- Btw. in many countries most urban sidewalks are "for pedestrians only" (i.e. highway=footway) but don't have any signs. It's a recurring misstatement that footway implies path+foot=designated, when it's in fact the other way round; things that are not path + foot=designated can also be highway=footway, but at least those that are path + foot=designated are equal to footway. And not all footways are equal to path + foot=designated. "Mainly or exclusively for pedestrians" is what matters the most. Alv (talk) 09:02, 31 July 2015 (UTC)
Now things have come full circle. As I stated above, my only motivation was to improve the distinction of footway vs path to give a crystal clear classification criteria. "Mainly or exclusively for pedestrians" was neither precise nor unmistakable. What was "mainly" supposed to mean? --geow (talk) 21:55, 31 July 2015 (UTC)
- Think of the time before highway=path hadn't even been proposed. Then we had to classify things as footway, or cycleway: if a way was allowed for bicycles (but not cars), and it looked and acted like a cycleway (see Duck tagging), it was best tagged as cycleway (which implied foot access was allowed, in most countries). If the way was only allowed for pedestrians, it was "exclusively" and clearly a footway. If the way allowed also cyclists, but it didn't look like it was built as a route to cycle on or mainly wasn't used as a cycling route (e.g. forest trails, narrow "paths" in parks, paths which wouldn't take a cyclist anywhere where they could continue their journey, paths with a bad surface to cycle on, paths with a much better alternative cycleway nearby), they were mostly tagged footway + bicycle=yes. (note "route" in this explanation doesn't refer to signposted cycling routes, but persons' trips. Mostly if the way had a traffic sign stating it was a cycleway (or combined cycleway and footway), it was tagged as a cycleway regardless of the condition. The path proposal never sought to change these meanings and the way they were used; or at least such changes to previous practice was rejected.
- Probably a global "crystal clear" classification criteria isn't even possible; it doesn't matter if the data user gets the relevant data either way. I presume you're mapping in Germany, which, as far as I know, has a rule that one may not cycle on paths that are narrower than n meters, and parks aren't sprinkled with footway traffic signs everywhere such paths cross a wider cyclepath? Consider the possibility that such rule didn't exist, i.e. one could legally cycle on those narrower ways, but mostly it was pointless, and if not going to, say, a park bench next to that narrow way, cyclists would stick to and prefer to describe only those "real" cycleways (even if they're combined/segregated cycleways and footways); the narrow ones would be an example of "mainly for pedestrians". Mappers have local knowledge, they can say if a way looks like a cycleway or not, even if there's no totally objective criteria (lacking traffic signs). We try to claim that the classification isn't based on "usability" by anything, but deep down it's built in the highway tag's meaning: if a primary road was left unmaintained for decades, but was still used, at some point the condition would be so bad that it would cease to be tagged as a road. Alv (talk) 11:47, 1 August 2015 (UTC)
- @RicoZ: I wouldn't even dream of that ;-) Actually - do we really need 5 highway types for non motorized traffic? --geow (talk) 18:01, 2 August 2015 (UTC)
My last edit from 21:16, 31. Jul. 2015 was reverted on 12:46, 2. Aug. 2015 by Mateusz Konieczny. He considerd my "redefinition to be a poor idea". So I proposed it at http://wiki.openstreetmap.org/wiki/Talk:Tag:highway%3Dfootway#Advanced_definition:_Distinction_footway_vs_path and I'm looking forward to your feedback Mateusz and you all. --geow (talk) 17:51, 2 August 2015 (UTC)
Or join the tagging list discussion at http://gis.19327.n5.nabble.com/highway-footway-Advanced-definition-Distinction-footway-vs-path-td5851506.html --geow (talk) 07:35, 3 August 2015 (UTC)
Your change to tag:highway=path
Hi, in this edit you have changed the semantics of the page substantially: what was
- "The default access restriction of highway=path is "open to all non-motorized vehicles (motor_vehicle=no), but emergency vehicles are allowed (emergency=destination)". (Although it depends on each country what vehicles are allowed by default)."
- "The default access restrictions of paths varies from country, see OSM_tags_for_routing/Access-Restrictions"
Now each country is "free" to change the default restrictions and indeed the table for Austria says that bicycles are prohibited on paths by default. However this is contrary to the approved proposal of highway=path . Has this been discussed on the mailing list or somewhere else? I am strongly in favor of reverting this also in the German page as it changes the semantics of a feature which has been in use for many years and may people and tools may rely on that. RicoZ (talk) 13:11, 8 July 2015 (UTC)
Hi there RicoZ, I changed nothing "substantially" by adding a link to OSM_tags_for_routing/Access-Restrictions. Contentwise this information was already included in the previous version (in brackets). Actually the legal access-restrictions are subject to constant global changing, therefore the link is needed. It provides more specified information than before and for many countries "open to all non-motorized vehicles" is still appropriate, no crucial semantics are changed. --geow (talk) 20:24, 8 July 2015 (UTC)
- Imho it is a very substantial change. The proposal was quite clear saying "open to all non-motorized vehicles and emergency vehicles". Now the text is saying that in Austria it is closed to all non-motorized vehicles. The text in brackets - whoever added it - could be interpreted in various other ways not contradicting the original proposal such as that in some countries some kind of forestry vehicles might be allowed in addition to all non-motorized vehicles and emergency vehicles. RicoZ (talk) 21:24, 8 July 2015 (UTC)
The essential path attribute "open to all non-motorized vehicles" is untouched by my edit. But this is only a general access rule. Actually from country to country this rule is restricted or extended for other means of transportation. I'm aware that Austria bans bicycles from paths and even tracks (!) in forestry lands by default if not otherwise permitted or signposted. On the other side of the globe i.e. in Nepal motorcycles are used on paths between mountain villages. Legal access restrictions should always be tagged properly on the individual way. --geow (talk) 17:35, 11 July 2015 (UTC)