My understanding is that the MIT license allows sublicensing while the BSD license does not, meaning that license restrictions can be added to things like mashups and not to the original data as a whole (this view taken largely from Larry Rosen's book).

Does it seem reasonable to state that these licenses provide a space between the CCBY and CC0 licenses, supporting the reuse of the latter with a subset of attribution requirements of the former?

Is there anything that is dramatically incorrect in this understanding? Anything which makes these licenses entirely unsuitable for data sets?

  • 2
    Don't forget about the Open Database License ... CC licenses were really not designed for data ... even John Wilbanks (former head of CC's Science Commons) advocated CC0 for data due to issues in using CC licenses for this type of information.
    – Joe
    Commented May 12, 2013 at 7:20
  • The ODbL aspires to be a copyleft license, and the OP only mentions non-copyleft licenses, so he probably is only interested in that. (Besides, the ODbL attempts to get the licensor to agree to terms there are no legal grounds for, and that only a fool would agree to. It is clearly the work of amateurs and have zero chance of standing up in court. I would personally never use the OBdL for anything. And I'd be quite happy to reuse data that somebody has tried to stick an ODbL on in any project without complying with the license.
    – user135
    Commented May 12, 2013 at 8:09
  • 1
    Open Data Commons also has an Attribution license and the Public Domain Dedication and License Commented May 13, 2013 at 10:26

1 Answer 1


A good reason to not use MIT and BSD licenses for data is that they were written for software, so they're not a great fit. And, CC-BY was written for creative works, not data.

I don't know where you got the impression that BSD doesn't allow sublicensing, but it certainly does. It also allows adding additional licenses, as long as they are license compatible.

CC-BY says that attribution needs to be given in the form specified by the licensor. If you want lighterweight attribution requirements, just say so in your CC-BY license.

Rather than choosing A, B, or C, you might want to start by writing down the goals that you're trying to achieve with your license. Given a set of goals, people might be able to suggest a license or family of licenses which is a good fit.

  • 1
    I got the idea that the BSD license doesn't allow sublicensing from rosenlaw.com/oslbook.htm, a book written by an IP attorney, and from conversations with Red Hat counsel Richard Fontana. Both suggest that the BSD license creates direct-to-consumer permissions granted by the author for his work, but that other copyright-worthy elements can be licensed in more restrictive terms, allowing the work as a whole, once significantly modified, to be under a more restrictive license. Commented Jun 2, 2013 at 2:26
  • 1
    This would be different from the MIT license which explicitly allows sublicensing, and therefore allows adding restrictions without creating a new copyright-worthy work. Commented Jun 2, 2013 at 2:28
  • Citation to a page in the book so I don't have to read all 13 chapters?
    – Tom Morris
    Commented Jun 2, 2013 at 18:13
  • 1
    You probably want to read all of Chapter 5, but the discussion of sublicensing and is under the MIT license section, and begins on page 87. It's worth noting that US case law has a presumption against licenses being sublicensable though this has been in overridden by statute for exclusive licenses. Non-exclusive licenses are, however, still presumed to be non-sublicensable. Commented Jun 3, 2013 at 1:09

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.