Potential Import of OCM locations into OSM
Now that OCM has over 27,000 electric car charging locations it would make sense to take a closer look at the options surrounding an import or bridge between OCM and OSM.
To this end I have updated the Open Charge Map page and I will post to the imports mail list to get opinions to see if there is general support for such an import.
In any case, OCM is now firmly established in the electric car arena and so it makes sense that there be some degree of coordination between OCM and OSM. I am hoping to trigger some discussion regarding the mapping of electric car charging stations in OSM. It is a new area and as of today there are only 4300 or so listed on OSM. I would like to see a project initiated within OSM related directly to electric car charging stations and this would of course tie in with any import discussions. Paul Churchley (talk) 11:33, 11 January 2015 (UTC)
Open Charge Map Locations
OpenChargeMap is a global database not UK specific as stated in the introduction. It contains 8823 Charging Stations (21 September 2012).
CC-BY-SA -> CC-BY ?
Could OCM consider releasing their data as CC-BY instead of CC-BY-SA? As OSM uses ODBL, it's impossible for us to share alike as Creative Commons, as ODBL will never be exactly the same as CC.
- I agree. I have asked OCM if they could relicense to CC-BY. Paul Churchley (talk) 11:27, 13 January 2015 (UTC)
I think if you include website, this url will already contain the ocm ref. In that case, there is no need for ref:ocm anymore. Sites like Openlinkmap can convert ref:ocm into urls of websites though, so just using ref:ocm instead of website would also work and be cleaner.
- I know that a ref:ocm=* could be converted to a url but if the station is looked at just through the normal mapping tools that url won't be obvious or available. So if people want to easily like between OSM and OCM then the url must also be tagged. I don't have any strong feelings either way to be honest. Providing the OCM ref is tagged then the link is created and can be followed by those that are interested. Paul Churchley (talk) 11:27, 13 January 2015 (UTC)
Import not Necessary – We Don't Need Your Data
I think that an import of OCM data (indepent which license OCM data has) is not necessary. As it is a automatical bulk import, we will get the same problems as every time and every place you do a bulk import. Bulk imports are often done by a very small group of people (or one single person). But because only this/these person(s) is interested in this type of data, the data won't be maintained very well (removing razed charging stations, change changed operators/networks etc.). If local mappers add the charging stations by themselves because they have seen them in field, they will maintain their stations. There are a couple of examples of badly maintained, imported data (e.g. US, The Netherlands, France) and, in opposite, countries where the data was collected manually, e.g. all German speaking countries. Mappers only care about the objects they have added theirselves to OSM, spending time on.
We had a special weekly task about charging stations at German OSM blog last year (German blog post, English summary). There are already > 1300 stations in DE+AT+CH! These objects have been added manually and checked by local people. If you import your data into OSM, you corrupt the quality of OSM data. OSM should be a hand-made high-quality crowdsource geodatabase and now automatically-mixed dataset like OpenAdresses!
Why don't you as OCM just declare "Hey OSM contributors, this is our database, you can look at it and add our data to OSM."? This would make a manual import possible and local mappers could decide if they import the data or not.
If somebody wants a combined dataset of OSM and OCM data, he needs compatible licenses but not one dataset because every data user can combine the two datasets by himself by just joining them and removing duplicated objects (e.g. looking at ref=*, operator=* and distance). --Nakaner (talk) 11:14, 18 January 2015 (UTC)
- After some discussion with Paul a good week ago, it became clear that they don't actually want to share their data, except maybe for the positions. They want to keep the monopoly over the details for each station though. On the one hand it seems like a good idea to store and keep data up-to-date in one central place, but on the other we'd prefer to have most pertinent details readily available in OSM anyway.
- Of course, there is no reason why Openstreetmap couldn't be that central place, but we'll probably keep 'competing'. If they do happen to change their license, we might be able to validate each other's details, but I don't see cooperation going much further than that.
- We could go and peek which of their data is licensed with a suitable license and then import that data directly from the original source, or go out and survey for ourselves. And OCM could use OSM data to improve their DB, as long as they add the source and that those tidbits are licensed as ODBL. Adding yet another license to the mix.
- Now, I'll go and check whether the handful of charging stations I know about around my city are mapped... the devil is in those details though, which is why I hadn't done so yet.--Polyglot (talk) 12:21, 18 January 2015 (UTC)
Stagnant discussion - how to move forward?
There seems to be consensus that OCM and OSM should not duplicate each others work, but there seems to be a roadblock every time it is discussed. It's been 5 years since this page has been updated and with the growth of EVs over the last couple of years, not to mention the next few, the need for a universal open source of charging station information is becoming more important by the day.
There is a discussion on the OpenChargeMap community site - I'd like to highlight a couple of comments by Christopher (OCM admin):
"On the plus side, OSM have at least 1000 times as many data contributors as we do and they will eventually all be driving electric cars, so maybe there’s hope in the long term. If an official from OSM said we want to take over Open Charge Map and make it part of the OSM dataset we’d probably go for it, as running OCM just cost us/me time and money :slight_smile:"
"Anyway, the root of the discussion is probably the question of whether OCM should exist at all if OSM exists, that’s one for the OSM community to answer by becoming the definitive reference for EV charging data, which would then make OCM no longer required. This is fine if it can be done, but it won’t be me doing it."
It seems there is consensus that long term, including the OCM administrator, that OSM will be the preferred data repository. I still think that OCM has an important role so far as an interface to the data, but the data itself doesn't need to be needlessly duplicated.
I find myself not being quick off the mark to update either OCM or OSM because I'm doing twice the work, and half the data will be trashed when they are eventually merged. It wouldn't be that bad if I knew which half :) Rather than all this talk about imports and licensing, could I suggest a very bold move:
- OCM stops accepting new entries to its database and acts as a gateway/interface to new entries on OSM. This reduces the future workload.
- OCM acts as a view or display for OSM charging station content. There are many other sites out there which do the same thing with other OSM data.
- OSM entries which already exist on OCM get a ref:ocm=#### tag so that at least there is a clear key that links entries duplicated across the two sites. [Note: This can be done independently of any other changes and I'm happy to make a start on it if people think it is useful]
I'm aware that the first two require a fair bit of coding effort on the OCM side of things. I didn't like to suggest it since I'm not a developer and it's basically me asking others to do what amounts to a fair bit of work. If Christopher agrees, then I expect any offers of assistance from experienced individuals would be welcomed!
I'll link to this post in the OCM thread and also the telegram group which was set up (which I joined and promptly forgot about, because I don't use telegram for anything else an just never opened the tab again!)
Update - just added ref:ocm tags on 3 locations in my home city - https://www.openstreetmap.org/changeset/92612140
- Thank you for moving this forward. I drive a Model 3 and try to map charging stations, when I can. When I first discovered OpenChargeMap, I expected to find the imperical data coming from OpenStreetMap. That is until I read Christopher's replies to that query you mention. I realized this would not happen until OpenStreetMap was seen by him as a reliable data import source. So.... with my limited time... I have ONLY entered chargers into OpenStreetMap and use OsmAnd to find them. Thinking, at some point, OpenStreetMap will have valuable data OpenChargeMap will want to import, and, will find a way to do it. --TreeStryder (talk) 14:59, 17 October 2020 (UTC)
- I'm also a electric car driver, and i'm contributing for chargers in Belgium (i'm also discussing with the private companies... to see if they could open their data :-) ). I have question on the idea of adding the "ref:ocm" : is it for an import ? Because if it is, it needs to follow the import procedure before adding it. We are strict about it in the community, and we don't want everyone adding their "reference" to their own database like that (as that's not the purpose of OSM). It also need to really be useful or needed ? Does it really prevent duplicate ? Is it only useful one time to check duplicate then it wouldn't be needed ? That's why the import procedure exist, to check if the tags are needed or useful (and accepted by the local community). It could be difficult to make that for every countries too. I would do the import process per country to make it easier and also because the status per country is quite different (as some countries are really well mapped, while other lack a lot of chargers).
- In general (similar to your suggestion), i would be on the side of just mapping things in OSM without special tags (using the official name and reference when it is open data) and then let OCM do its own filtering if they want to do it (like any application user can do it) - or just use OSM like it is without any other source (they do what they want :p). I would also support the idea that they switch to contributing directly to OSM as that would remove the duplicate.
- One important point to always remember : OSM is a database for any usage, so we shouldn't dedicate it to one service only (especially as we don't know if it will still exist in 5 years, OsmAnd could become the go-to EV charger app in the future nobody knows). --Anakil (talk) 13:05, 18 October 2020 (UTC)