So, looking at geonames, I am finding that there are a significant number of blindspots regarding zipcode mappings. Many countries are missing. Some countries have abbreviated data (prefixes in zipcodes only are listed in Canada for 'copyright reasons').

So my current approach is to create a routine in the software (an open source program called LedgerSMB) which will let users export a list of country/state or province/city/mail_code mappings to a CSV file that can then be emailed back to us for processing. No identifiable information would be included and the results would be merged before distributing. My current checklist is:

  1. An agreement with the customer that this will be going into a very permissively licensed package in exchange for validation feedback on their data.

  2. We'd offer attribution in an appropriate form (but no downstream guarantees).

  3. Data would be run through automated checks, and normalized prior to merging.

Is there anything major I am missing? Anything we need to think about before approaching users (many being business users who are not entirely technical) about this?

1 Answer 1


Do you believe that these data can be licensed? If so, then that would imply that you should ask the user to verify that they are authorized to accept your licensing terms.

  • Agreed -- the original question puts scare quotes around "copyright reasons," but Crown copyright is a real concern in Commonwealth countries, as is the potential for copyrighting facts (which are generally not considered copyrightable in the US). There's a reason that data isn't already available, and I'm not sure why new users would be any more empowered than existing researchers to collect/redistribute it. Commented May 17, 2013 at 20:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.