Some schools are well mapped, with many relevant features tagged. Pretty much any object than can be mapped elsewhere, can be mapped within a school, with one exception: toilets. Because (at least in some countries) school grounds tend to be private access, a strict reading of the amenity=toilets tag’s rules would mean that these can’t be tagged as such, because access=private is deprecated for toilets.
This is inconsistent and may also produce problems for outreach, when students legitimately mapping their own school get reverted.
Aside from lack of renderer support for hiding access=private toilets, why is this a rule in the first place? Could we come up with a more consistent policy?
I think the problem with schools is that they have an access status that actually falls between access=customers and access=private. The trouble is that access=customers generally implies a fairly permissive access.
The reason for not normally mapping access=private toilets is that amount of noise they would add to the map and the risk that data consumers would fail to detect that they are private. That also applies to cafe and restaurant toilets, which may be access=customers by law, but should be findable without a map by those close enough to use them and allowed to use them.
I would actually suggest that school toilets are no different from many other features that might get mapped within a school.
I’d also suggest that there are other places where it might be appropriate to map private ones, because the building is actually available for public hire (as in fact are most schools).
I’d certainly suggest that changesets deleting these should be challenged, and possibly escalated to the Data Working Group.
Most rules in OSM should be subject to common sense.
Actually, school toilets generally have additional access restrictions on toilets. Some will be staff only and some children only.
Toilets in buildings should just not be mapped using amenity=toilets except in a few cases where the buildings are publicly accessible throughout the day (for instance many universities, shopping centres/malls etc).
On the other hand for effective indoor mapping another tag needs to be used for the examples given by hadw. For instance a number of schools also double as sports facilities or community centres out of school hours, so some of the school building is fairly accessible. However, I would tend to avoid the amenity=toilets tag in these cases. Remember also that many tags will come with in-built assumptions: and the notion that sports centres don’t come with changing facilities and toilets wll be one of those.
In general toilets=customers as a tag is fine on pubs/restaurants etc. It is important to map availability (and if possible suitability) of disabled facilities in such places.
However the problem is with definition. Current tagging schemes don’t allow us to distinguish standalone toilets from the toilets being internal infrastructure. I have the intuition that standalone toilets buildings and paid toilets on the train station are similar and deserve amenity=toilets tagging, but what if they are part of the shopping centre (which is as public as the train station), school, office building? What if they are not paid? Does it make sense to add just an indoor tagging without amenity - it should be still tagged only for public spaces, not private buildings, so where is the difference? It’s hard for me to draw the line.
That’s why I want to demote toilets rendering on osm-carto soon:
What about toilets in supermarkets? Of the 2.5 supermarkets in town (two trading, one gone bust) in town, only one has customer toilets. I haven’t mapped the toilets yet, partly because I didn’t know the best way to do it.
One thought I had was to map it as a node inside the supermarket building, and have amenity=toilets and access=customers. It’s not room-detail mapping but probably good enough (especially for men, whose aim is notoriously bad).
I think I’ll wait until some sort of consensus is reached.