Visiting a National Trust property at easter? Here's some useful mapping

Communities United Kingdom
1

There seems to have been a bit of a misunderstanding around gate mapping in some National Trust areas in England, Wales and Northern Ireland. Specifically, there seems to have been a bit of a thing of mapping “values for multiple barrier features in one node”. This overpass query shows a selection (the 174 with the nonsensical value locked=yes;no set). An example is https://www.openstreetmap.org/node/2346780305:

barrier   swing_gate;kissing_gate
locked    yes;no
material  wood
width     3.60;1.12

All of those values (apart from wood) seem unlikely. My recollection is that that one was just a regular field gate; my guess is that a kissing gate has been added in the last few years. It’d be good to know how wide the gate that wasn’t locked is, and that needs each gate mapped separately.

Most of these edits seem to be by NT people in the field using Vespucci. I’ve mentioned this one on the changeset to the NT person who made it, but as there are lots more I suspect that it could benefit from more eyes cast on the problem.

I suspect that a few people will be visiting NT areas over the Easter weekend, so thought I’d mention it here too…

1 Like
2

Might be worth taking it up with the NT person who spoke at SOTM Dundee. They were recording their sites on OSM as a centrally managed project so this might be best handled in a similar manner rather than the individual mapper.

https://2025.stateofthemap.eu/#session/9EKBSB/

3

Is this the gate(s) in question?

The photograph is of the point where a bridleway and the track it has been following diverge. The bridleway passes through the gate on the left and into the field. The way itself connects the park entrance near [[583842]] (behind and left of the...

4

Trying again by email as the forum hid the URLhttps://www.geograph.org.uk/photo/5115502
The photograph is looking eastward along a bridleway that connects the park entrance near [[583842]] (behind the viewpoint,) with Ault Hucknall (ahead). The gate ahead (which really is at an angle) is the point where the bridleway and the track it...

5

It’s likely to be, although that’s from 10 years ago. A bit clearer is:

5115502_aa4b557c
5115502_aa4b557c640×480 230 KB

( © Copyright David Dixon and licensed for reuse under this Creative Commons Licence.)

That’s looking west; personally I’d suggest that two gate nodes needed to be mapped there to capture information about both gates. The original mapper was me 13 years ago; I likely just ignored the locked gate at the time.

It still would benefit from checking as I’d map neither of those gates as a swing_gate or a kissing_gate - but of course either might have been replaced in the last 10 years.

The other picture is from the other side and is at Geograph here.

6

It isn’t very clear what the work flow was that led to this mix up, if anybody can get hold of somebody that can explain that it would be interesting to know (it is not really something you can do by accident).

7

I’m guessing it’ll need info from the people doing the editing, and I suspect they (as NT volunteers) might be a bit busy this weekend and next week until the schools go back, at least.

1 Like
8

I think the small pedestrian gate on the left in the photo does have a kissing-gate setup, going between two ends of a fence enclosing a small region between the two sides of the main barrier. If so, I would map that as a kissing_gate.

For the swing_gate, I’ve seen NT mappers map these standard vehicle field gates as that elsewhere, but haven’t had a chance to follow up with them. I think it’s probably a misunderstanding/error in some internal documentation their using. I have a contact at NT I can get in touch with about that.

I’m not sure the style of combing two parallel barriers on the same path/track crossing the same linear barrier is necessarily wrong. If we would consider the route to be a single track leading up to the fence on both sides, and then there are two parallel ways to pass through the fence, it’s a bit like the question of whether to split a highway at a small traffic island. You could conceivably map a parallel pair of node barrier values like they’ve done with semi-colon lists. But on balance I think it would be better to split the approach ways and map the node-barriers as two separate nodes, not least because it makes it much easier for / more likely that data users will interpret the situation correctly.

1 Like
9

Well multiple values in a toplevel tag is essentially unsupported and for good reasons. So while there might not be any information loss in the tagging as such, for all practical purposes nobody is going to be using the data.

2 Likes
10

I normally create a short section of footpath to go through the parallel pedestrian gate or stile with appropriate access tags on the section through the big gate.

Sometimes the big gate is for farm access, sometimes a stile is provided on bridleways to save walkers opening the gate.

5 Likes
11

In Sheffield there are quite a few places with there are three barriers in one: OpenStreetMap

Where there is a swing gate for vehicle access, a horse stile for horses and a motorcycle barrier for pedestrians and people using standard cycles. I’m not sure I like how I’ve tagged them but I also think it’s the least bad way.

3 Likes
12

Hi all,

Thanks for flagging the issue with mapping multiple barrier features into one node, and bringing the One feature, one OSM element - OpenStreetMap Wiki article to our attention.

In the original example, we have mapped the adjacent features together, with associated values in corresponding order (ie. the swing gate is locked, wood and 3.60m wide, while the adjacent kissing gate is not usually locked, wood and 1.12m wide).

The logic behind mapping the features together was such that it isnt always the case that there is a separate path and track for each barrier to be joined to.

That said, we have noted that these combined features are difficult to symbolise in maps, and can appreciate that this may present challenges for routing. We are therefore happy to incorporate a review of the combined barriers as part of our systematic QC of the NT promoted walking / running trails that we have surveyed in the last year.

We will also reach out to the local property to confirm whether this gate should in fact be mapped as barrier=gate rather than swing_gate.

For full transparency, as part of our QC we are also seeking to:

  1. Check the tagging on the NT promoted walking / running trails (relations) for spelling mistakes, missing tags, & incorrect formatting.
  2. Remove the car parks from the relations (they are linear they are contributing to our trail lengths).
  3. Add roles where appropriate (approach, excursion, alternative) to support routing.
  4. Review spelling mistakes, missing tags & incorrect formatting on the ways - focusing on the surface, smoothness & width tags that we have surveyed along these routes.
  5. We are also aware that some of the relations have several forms of address in the from & to tags as they have been lifted from our website (ie. Some include grid references / what3words rather than just postal addresses) so we will seek to do a review of these tags too.
  6. Lastly, there has also been some discussion after the SOTM conference in Dundee around the need to remove point features from the relations as they are superfluous for routing.

With that in mind, please bare with us as it will take some months to complete our QC (including a review of multiple features mapped as one node) for all of the properties in our care across London & the South East, South West, Midlands & East of England, the North, Wales & Northern Ireland.

We are grateful for the feedback and keen to learn from the wealth of knowledge that the OSM community has to offer, so thank you again for flagging.

Kind regards,

National Trust GIS Paths & Trails Team.

4 Likes