A question about `railway=site` in the UK

I’ve notice that there seem to be quite a few railway=site in the UK, and that’s something that’s not documented in the OSM wiki. It is documented in the OpenRailwayMap wiki as something that doesn’t really occur in the UK.

What usage there is seems to correspond to former railway infrastructure - other keys such as historic:railway=junction and historic:railway=station. Excluding historic keys there is very little left - these seem all to correspond to timepoint locations (see the OSM wiki for that and externally).

I suspect what might be happening is the a well-known “historic railway mapper” is exploiting the fact that Nominatim indexes railway keys (for example a search for “Sinnington” finds the “railway site” before the village despite there being no current railway infrastructure there).

Have I missed some meaning that railway=site is conveying on these historic stations and junctions? If not, I propose to clean up at least the ones locally to me, where I know there is definitely no current railway. I’m only suggesting removing railway=site (i.e. leaving any historic:station tag), and I’m only suggesting removing it where the usage is historic. There are no advertised consumers of this data.

Edit: Also mentioned in this German thread.

I think it’s more likely that railway=site is being added to make a label appear in OpenRailwayMap:

e.g. the many disused sites in County Durham:
image

[I must confess to having added some of these myself, having copied tagging use from elsewhere.]

It is a tagging-for-the-renderer hack, and ORM should render historic: versions if it does actually want to show historic sites. but personally I don’t think it is doing any harm.

1 Like