To get a complete example I would suggest adding the following tags to the relation:
authentication:membership_card=yes
authentication:none=no (need some form of authentication even if it is free)
maxstay=unlimited/1 hour/30 minutes
motorcar=yes (no e-bike charging)
parking:fee=yes/no
payment:app=yes/no
payment:cards=yes/no
payment:cash=yes/no
payment:membership_card=yes/no
website= (at least for the network or better if there is a link to the station itself)
Thanks for the analysis. Can you run another test on the single node ones, to see how many come back with capacity= 1 or 2 ; capacity>2 and without capacity?
At least here in Germany there are still quite a few installations that only have a single charger (we are not quite as far as Norway yetđ ). I can think of 3 different supermarket parking lots that only offer one.
Thank you for the info. Iâve added the tags which I know the values for. I do not know the output voltage or amperage of the chargers, so I didnât add that.
Thank you for producing the draft, @jmarchon. It makes it easier to discuss the details.
Main comments:
I think amenity=charging_stall is not a good choice of words. The use of amenity for both the charging site and the charge point will blur the intended hierarchy and will create confusion. Also, stall is ambiguous and perhaps not well understood across languages. I think your original proposal of man_made=charge_point (or man_made=charging_point) is better, and it also resembles the mapping of man_made=pump for fuel stations.
The proposal should make it clear that mapping charge points is optional (the text says âshould useâ now).
Ways and relations: The relation site concept is too complicated now. I think we should rather reuse the concept of power=plant / power=generator. Look at the mapping description at Tag:power=plant - OpenStreetMap Wiki.
If the charge points are grouped close together then amenity=charging_station is put on a closed way encompassing that area.
If the charge points are dispersed then amenity=charging_station is put on a site relation with all charge points as members (or if the user prefer a relation instead of a closed way). The relations get a type=site tag and otherwise the same tags as for the other two cases.
If there is only one charge point then amenity=charging_station is a single node. The charge point is not mapped in addition. A node in the middle of the charge points could also be used if no way or relation is mapped.
This way/relation structure works well for wind farms and is being used extensively. A benefit of the above scheme is that the same more detailed tags (subkeys) are used in all three cases. Also, there will be two features instead of three, and there is flexibility for the user in the choice of node, way or relation based mapping.
I think we should not include all kinds of other features into the relation (parking, cabinets etc). It will just complicate things. Simplicity is key for a successful tagging scheme, I think.
For debate: We could perhaps mention that man_made=charge_point could be used alone, without an associated amenity=charging_station. The use case would be single chargers for example outside an office, which are not really a pubic destination for charging (like a fuel station), or widely dispersed poles along a street.
I would recommend that we do not redefine the subkeys (socket etc) in this proposal but just refers back to the original wiki page. If you insert the {{Tag|key|value}} template also in the paragraphs it will be easer for the reader.
I would also recommend that we keep the list of subkeys to a bare minimum in order to avoid sidetracking the discussion and to increase the likelihood of approval. Every key mentioned could lead to a debate. I would just include the following keys:
name - for the site
ref - for the charge point
brand - for a branded site
operator - for a non-branded site
capacity - for both site and point
socket: - for both
socket::output - for both (with a note that it denotes the highest output available to on a site to the consumer)
network is not well defined. charge is very difficult to keep updated (prices often vary on a daily basis). current and voltage are not needed for anything (output in kW contains what is useful for the consumer and is the number being used by the operators in their communication).
Replace the technical term EVSE with a normal word (what is it?).
An extra section on the proposal page with a few real world examples in a table would be helpful and convincing (a picture + tagging). Cases: The three node/way/relation cases above + one case with charge points along a street.
Including the statistics I provided could be helpful for avoiding the perception that charge points rather than sites are the most common feature mapped today. Just the number of sites vs points. Also not that the proposal is backwards compatible regarding subkeys.
Sure, of the single charger sites there are 4658 with capacity=1, 30454 with capacity=2 and 11205 with capacity=3 and higher.
In my country such sites tend to get expanded as the penetration of electric cars increases.
Itâs a good move forward to go on with a proposal, thank you.
According to examples here, I respectably disagree.
OSM would take risks to redefine what actually exists in other standards. A station is not a site, a pool is.
The station is the machine where you pay on, with one or more service equipment attached, with one or more sockets on each.
Donât redefine what a station, a pool, a service equipment is please.
At some point, it will be impossible to make third party data match with OSM and it would be a pity for maintenance with open data.
Chargers have a current limit and a power limit, wichever comes first will be the limitation. You can get 500A with a 400V car giving you 200kW. But you cannot get 500A with a 950V car, it will be limited to around 370A.
In plain English, itâs a âcharging stationâ, as a parallel to âgas stationâ/âpetrol stationâ. Nobody besides industry insiders calls it a âcharging poolâ.
Current and voltage are very important for end users.
If you drive a 400V car (the overwhelming majority) and arrive at a 180kW charger that is limited to 200A you will only get 80kW.
Thank you for your feedback, but I disagree. Here, when someone says âcharging stationâ, they almost always mean what you call the âcharging poolâ.
The data presented by @NKA earlier in this conversation supports this, I would say.
Itâs not me but the whole industry and many drivers actually.
Iâm not responsible of the wording here, someone already define it way better than me.
Note that âa poolâ translates to âune stationâ and âa stationâ to âune borneâ in French. Many people may be confused by that.