Extracting OSM Place Name Data to Generate a Dictionary Database

We are plan to extracting and using OSM name data (place name etc…) to generate a dictionary database. Is it correct to understand that the following usage does not mean creation a “Derivative Database”?

  1. Extract “name” from OpenStreetMap database using SQL query;
  2. If necessary, modify some of the character strings in the “name” (for cleaning and correcting spelling errors etc…);
  3. Binaryize the “name” data according to the unique dictionary format created by us; and
  4. Built the database into the customer’s on-premises environment, and allow access and view only in local environment (not published online).

It is probably not a good idea to ask for legal advice on the Internet particularly considering that providing such is heavily regulated. Therefore the following should simply be taken as mindless rambling.

An extract of OSM data is essentially always (the exception being insubstantial extracts) a Derivative Database. The sharealike terms simply only kick in once you are using it publicly.

If you are doing this work on contract for your customer and not yourself then that is unlikely to be publicly used. If you are selling it as a product then that would be different, but in both cases you wouldn’t be able to restrict how the customer uses the data. You should in any case include a copy of the licence with the data as your customer might change their mind and will need to know on what terms they can use the data publicly.

IANAL and in particular not yours, if in doubt you should ask your own counsel.