Having a choice of licenses at dataset create/update time by default encourages bad practice. It's really bad for data reusers if all the datasets on your site use a lot of different licenses.
Open Knowledge's own advice (I think this is written down on somewhere like opendefinition.org or opendatacommons.org) is to pick one license and use it throughout your site.
So we should pick one license to be the default for CKAN and hide all the others.
I'd suggest that we choose the ODbL as recommended by opendatacommons.org.
Of course we can't force open data publishers to use this license and most CKAN site maintainers will probably customize this to use their own license itself. But at least we'll be encouraging good practices by default: Ideally use this one license for everything, or customize it to use your own license instead but at least use one consistent license. The current defaults encourage almost the worst practice.
@jpmckinney said:
If we have to pick one, I would pick CC-BY-4.0, which is increasingly used and recommended, e.g. this document commissioned by Ireland. Open Data Commons licenses, in general, are far less used than Creative Commons.
However, since a debate about licenses can last a long while :), I wonder how difficult it would be to add a step for choosing a "first license" as part of Getting Started.
I said:
I quite like Sean's idea to push one licence, but there still needs to be a way to select other licences. Having a default value pre-selected is probably a powerful-enough thing. Whilst a CKAN for a city will usually get to choose its licence for all its datasets, and forcing a single licence might well be suitable, it's also quite common (e.g. London Datastore, data.gov.uk, data.gov, new EU-wide aggregation of national portals later this year) to see CKAN used as an aggregation of a number of data owners, each with different licences, so you do want easy selection. data.gov.uk would not want a default value pre-selected, since we'd rather publishers made a conscious decision to select the correct licence, rather than get it wrong. So ideally we'd just have a config option containing the default licence, that data.gov.uk would change to blank.
(Conversation moved from https://github.com/ckan/ckan/pull/2002 )
@seanh said:
@jpmckinney said:
I said: