Hi folks, I’ve been looking into Fife’s core paths (pubished around 2012-2014) some more and intend to add them to OSM with designation=core_path, along with foot=designated, bicycle=designated, and horse=designated (per law as in Land Reform (Scotland) Act 2003, s 17(2)(a)). I’d be doing this by using the core paths data as on overlay and then manually drawing/tagging highways in OSM rather than importing the data directly, but first, I’d like the community’s opinion on:
Core Path Accuracy (not the GPS kind): there’s a few cases I’ve seen where the drawn core paths–as provided by Fife Council–don’t make any physical sense, and checking old photos of these areas, the paths never made sense. For example:
This shows a core path in Dalgety Bay; the important bit is the segment that crosses the road. Namely, the position of this crossing is nonsensicle as there has never been a crossing in this location. I would normally assume this to be a simple GPS error or something, but the crossing appears to be west of the northward junction (with high kerbs), even though there was (and is) a perfectly usable lit crossing just east of the junction. Altogether, it makes it sound like this was maybe intentional? Here is a street view photo of the area in question.
The purple line is the core path as provided by Fife Council; the pink line is the highway that I would prefer to designate a core path as it makes more sense. Here’s the OSM link to the area in question.
Normally, I wouldn’t fuss so much about an academic tagging exercise, but core paths do afford a tangible legal benefit so I’d like to get this correct the first time. The question here is: would it be reasonable to assume that Fife Council made a mistake in drawing the core path across the road where they did (~30m from the lit crossing)?
I would say yes but really we will never know unless you can speak to the person who digitised them. I suspect they traced out routes without checking all the details first.
See also similar question about English “definitive” data here.
Obviously the legal rights (especially in open country) in Scotland are very different, but routes drawn by humans may still have been drawn in the wrong place by mistake.
That would be my suspicion too. I’ll try to find and contact the relevant team and see if I can get something sorted.
And looking into the core paths further, I’ve found 4 more disrepancies in Dalgety Bay alone which can’t be explained by a simple misalignment. This might be more extensive than expected . Oh well, no-one said it would be simple to ensure OSM is the best map there can ever be
I’ll keep this thread updated if something comes up; core path mapping in much of Scotland appears to be extremely sparse so far, and anything I get from this experience should be applicable to the other council areas.
Is it possible that the path route was set before the crossing was in existence?
I’m not sure about Scotland, but in England and Wales, Public Rights of Way that go along pavements through built-up areas are typically the result of a diversion of an historic route that existed before the development. In the planning phase, the developer would propose a route for the path to be diverted following the new planned roads. In the example above, if the crossing wasn’t there, then it would be more sensible to cross as the route does, since that means you only cross one road rather than two.
Again I don’t know about the situation in Scotland, but in England and Wales if a route is draw on the Definitive Map, then that is a legal route (unless or until it gets diverted), even if there’s a mistake in it. That doesn’t apply to unofficial digitised working copies though, so if the mistake has been made transferring the route from the legal Definitive Map to a digitized version, then it’s the Definitive Map that decides the route.
If the legal route does indeed go as shown, then the core_path part of it should be mapped as a way on that crossing, but not necessarily with a highway=footway tag. I guess it depends whether you think anyone would ever want to cross there. Looking at the map, I think it’s possible that they would. If you’re coming from the north and want to go down the first side-road to the south, it looks like the fence stops you crossing sooner, and you might choose not to make the detour to the crossing and back. So possibly I’d map the non-crossing as highway=footway + footway=link with a crossing=informal on the intersection node.
I emailed the Outdoor Access team for Fife with my queries and got a response back in a few days. Unfortunately, it doesn’t really answer any questions; only that the the Core Path network hasn’t been reviewed since its inception in 2012 and that core paths will be surveyed when time allows. Ah well.
I’ll add the core paths using my best judgement, though one question remains before I do:
I’ve a segment of core path where it is fairly apparent where it used to be, but is completely impassable due to gorse at points (all year round). Even though the path doesn’t currently exist in reality, would it be reasonable to add it anyway with access=no given its status as a core path?
I’d hoped that the Scottish Outdoor Access Code would have covered this, but the language in that document is somewhat vague. I do know that in Scotland you’ve got a more general right of access that I have down in England, but I don’t know what extra access rights you’ve got on a Core Path, even a blocked one. If you don’t own the land, is it legal for you to get your secateurs out as it would be in England and Wales on a public footpath?
I suspect that the bit that’s dubious is not “what the access rights are” but “whether it is any real sense any sort of highway in OSM terms” (even though the access authority has labelled it a Core Path), but I’d very much ask what people locally in Scotland are doing rather than assume that an approach that works in the very different regime south of the border is also a good fit for Scotland.
As you can see it’s relatively sparse. I think the reason for this is that there doesn’t tend to be any signage that indicates core paths and therefore we can only tag them by referring to official sources.
Where core paths are useful, especially in built up area’s a core path explicitly does allow cycling, and this means that in ExoMal’s example at the top the core path pavements can be cycled along. So knowing where the core paths are is useful for tagging access. I think in cases where the route to follow deviates a small distance from the “official” core path, there is some leeway. However I would probably just leave a gap in the core path mapping or map the core path as it’s defined without a highway where it deviates,
Largely the same thing applies where it is impassable, map the designated rougte, but ensure the tagging indicates that it can’t be followed. Worth also emailing the council access team about these, but it doesn’t sound like they are going to do much about opening these or clearing them….
At the moment, I think I’m inclined to follow @Chris_Fleming’s suggestions in general, so I intend to:
Map the core paths in OSM as they appear in the official data (adding designation=core_path).
Add bicycle=yes where not otherwise implied (e.g. steps, sidewalks) (I guess horse=yes should be included too, but I have no idea if data consumers interpret horse tags differently to bicycle tags so won’t touch that for now).
Allow a little leeway i.e. if the core path is slightly off from the actual highway, use reasonable judgement to identify the highway the core path is following.
In cases where the core path is blocked but could potentially be revived (e.g. overgrown gorse), then map the highway with access=no and any other relevant information (e.g. description and obstacle tags.
In cases where the core path is known to be diverted, then only the diversion should be mapped as a core path (stopped up core paths with an alternative diversion legally cease to be core paths).
In cases where the core path is blocked but cannot generally be revived (e.g. new buildings), then the core path should not be mapped at all (I’m not entirely sure about this point, as any such path should surely have a valid diversion by law so would fall into the above bullet point anyway?)
One day it would be pretty sweet to run a map matching algorithm to see how the OSM data differs from official core path data, like how the Walk Wheel Cycle Trust did with the National Cycle Network routes.
. Oh well, no-one said it would be simple to ensure OSM is the best map there can ever be 
