Is it OK to link to external datasets on incompatible licences?

INSPIRE dataset apparently cannot be imported into OSM (or at least parts could not be in 2013, see INSPIRE - OpenStreetMap Wiki )

It is still OK to link to specific entries using say ref:inspire, right? In the same way as we can put website links leading to fully copyrighted websites?

I am asking as I want to add Linking using {{tag|ref:inspire}} to external entries remains OK and can be useful. note to INSPIRE - OpenStreetMap Wiki page

(there is a national database of historic monuments in Poland with some weird and problematic identifiers - for some time they started assigning also INSPIRE codes which are far less problematic)

warning: I am not a lawyer. This message is not some official statement of any organization. Like any other post made by me (unless clearly stating otherwise).


Yes, of course it is ok.


This isn’t really a place to ask for legal advice. If you’re wondering if it’s a good idea, maybe it would help to know how readily usable the resulting data would be, and if it duplicates any other dataset.

Wikidata has a well-used INSPIRE ID (P4115) property, so wikidata=* tags can be an alternative. However, it would still be nice to tag it in OSM if it tends to appear on the kinds of map features that don’t have Wikidata items yet, or if INSPIRE is particularly commonly used in mapping workflows.

1 Like

Let me try to sum it up:
I have read several statements in the OSM space about INSPIRE data. They testify to doubts about whether INSPIRE data can be imported into OSM.
For example, at the link provided by Mateusz, until recently one could read about INSPIRE:
“The data is on licensing rules making it unusable in OSM.”
This has recently been removed. See discussion → Talk:INSPIRE - OpenStreetMap Wiki

Here is another thread testifying to these doubts:

I understand that:
There is no single INSPIRE database. There are many INSPIRE databases run by different institutions in different countries. Perhaps each of these databases is a different case.

Mateusz’ question refers to a specific database with information about heritage objects in Poland:,rejestr-zabytkow-nieruchomych/resource/53559/table?page=1&per_page=20&q=&sort=

The thing is that INSPIRE IDs are practically the only ones that uniquely identify heritage objects in Poland.
But this database is based on a licence: CC BY 4.0
Which means:
I can do practically anything with the data but it is necessary to: “Attribution - You must label the work appropriately, provide a link to the licence and indicate if changes have been made to it. You may do this in any reasonable way, as long as it does not suggest endorsement by the licensor of you or the way you use the work.”
[Such a licence - for a government institution maintained with my taxes seems like an abuse - but that’s another story.]

We only want to take the INSPIRE ID from the database and copy it to ref:inspire
I understand that this does not violate the licence.

As Mateusz wrote:

Direct ref:inspire value is more useful than wikidata as you can use INSPIRE ID as reference and get more info about a heritage object from,rejestr-zabytkow-nieruchomych/resource/53559/table?page=1&per_page=20&q=&sort=

If wikidata publishes INSPIRE IDs, it is an additional argument that this can also be done in OSM.

1 Like

Makes sense. Normally, Wikidata would also link to a page like that based on the property value, and you can even use Hub to automatically redirect from, say, an OSM ID to that page. But unusually the INSPIRE ID property doesn’t link anywhere. Is that because there’s no single website (even unofficially) that aggregates all the INSPIRE datasets?

I am not a lawyer, but this is how a lawyer (in the US) explained it to me: Facts cannot be copyrighted, but, an ID that an organization assigns to its data isn’t a “fact” because no one can generate such an ID independent of the organization that assigned it. An organization could claim that their ID is their “secret sauce.” Would that hold up in court? Who knows, but who wants to risk it? It also suggested that the data may have been copied from the source in question. Of course, the law may be different outside of the US.


Sorry if I’ve missed something - don’t most external datasets we link to, including Wikidata, have incompatible licences?

I’ve always taken that to mean we can’t copy over the contents of the dataset itself to OSM.

Are you concerned that the mere act of accessing the external database to identify the matching record, and then copying over the ID into OSM, could infringe their database right?

If that was the case, wouldn’t all external IDs be problematic?

I suppose you could ask the LWG if you really wanted peace of mind.

1 Like

So we believe that the datasource has a license that is not compatible with OSM, and therefore it is not permissible to copy from it… except for this one field? Why is one field in the database different than any other?

1 Like

Because this is ID in this database. It is unique. Every record has a different ID. ID uniquely identifies records. When you know ID, you can get its record.

EDIT: INSPIRE_ID is like URL address of a website. If you know URL, you can open a website. Website content can be copyrighted, but URL address is not. INSPIRE_ID is a similar case. Almost literally - see (b) example below.

For example if INSPIRE_ID is
then you can get its record.
For example:
You can search / filter this database:,rejestr-zabytkow-nieruchomych/resource/53559/table?page=1&per_page=20&q=&sort=
You can construct URL:,rejestr-zabytkow-nieruchomych/resource/53559/table?page=1&per_page=20&q=col1:%22PL.1.9.ZIPOZ.NID_N_12_BK.206051%22&sort=
and it directs you to specific record.
You can also quickly search object using this ID on this map:

There is no single INSPIRE database. There are many INSPIRE databases run by different institutions in different countries. Many databases.
Unofficial single giant INSPIRE database? I don’t think so.

I’m not an expert but this is what I know:

INSPIRE is a European standard.

See INSPIRE ID examples (from the herigage objects in the Poland dataset mentioned several times above):

The first part of INSPIRE ID is namespace / data source.
This is PL.1.9.ZIPOZ or PL.1.9.ZIPOZ.NID_N
PL stands for Poland.
The second part is a local identifier which is unique within the namespace / data source.

See also the explanation here:

Now see natural monuments in Poland (trees) INSPIRE IDs (I found them among OSM tags):

PL stands for Poland.
My guess is that this data source is PL.ZIPOP.1393.PP
This is a different data source than the heritage / monuments dataset (PL.1.9.ZIPOZ.NID_N).

There is not a single INSPIRE database. There are many INSPIRE databases run by different institutions in different countries.

I tried to find the list of these databases.

What I found is:

For example:

  • Open the URL above
    and select in sequence:
  • High-Value Datasets
  • Earth Observation and Environment - Show More
  • Protected sites
  • Poland
  • Rejestr zabytków nieruchomych [it means: Register of immovable monuments / Register of immovable heritage objects]

And you get info about the database we are talking about:

This page contains some info about it, but it seems to contain neither a download option nor the URL of the published dataset.

Even though INSPIRE is a European standard, a unified ref:inspire=* key seems inconvenient to use because there isn’t an obvious way to go from the ID to the webpage with more information. But we already have that problem with ref:vatin=*, another European standard.

In Wikidata, if a source of external identifiers lacks a coherent URL pattern, an alternative would be to add a URL (P2699) qualifier to each individual statement. Some properties, like OpenStreetMap numeric user ID (P8754) enforce this constraint. If each dataset has record pages at predictable URLs, then a less tedious alternative would be to create an item for each dataset and tag it with a general property-based URL formatter (P8183).

Interesting, I was unaware that database rights didn’t apply to unique fields.

Interesting, I was unaware that database rights didn’t apply to unique fields.

the question with wikidata is different because it is published with a cc0 license, but the OpenStreetMap-Foundation believes there might be third party rights in the data which haven’t been cleared so importing into OpenStreetMap is frowned upon. There is 0% probability that wikidata ids cannot be recorded because they are protected.

1 Like

It almost humorous how a community commenter notes to have made a wikidata code to then enter the created code into an OSM entry. Just saying…

Thanks for the explanation. That makes sense for Wikidata, but my understanding is the question is more general.

We link to external datasets all the time that are not openly licensed. Facebook and Twitter are the two big ones.

In general a list of IDs is going to attract very little protection when those IDs are computer-generated or otherwise involve nothing new in assembling them.

Where I would be careful is the normal issues for external identifiers. I would only consider it for something like a POI where it is unlikely to be merged with other features or split. It’s not great for landuse.


I am not aware about statements made by OpenStreetMap-Foundation on this topic or by LWG or by anything official. Though maybe I missed something?

(I made some comments worrying about this aspect, but these were not official OSMF statements, in fact I made nearly or all of them before being elected)

1 Like

I am not aware about statements made by OpenStreetMap-Foundation on this topic or by LWG or by anything official. Though maybe I missed something?

I was referring to if the information given there is inaccurate it should be updated.

I am not claiming that this advice is inaccurate, just that it is - AFAIK - not an official OSMF statement.