[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Stack Overflow and copyrightability of small snippets

On 2020/04/23 03:47 AM, Rebecca N. Palmer wrote:
Many packages include code snippets from, or based on, Stack Overflow answers [0].

Stack Overflow user-posted content is under CC-BY-SA (the version depending on its age) [1], which is a libre license but usually not the license these packages claim to be under.  Also, attribution is usually provided as a link only, not the longer form Stack Overflow say they require [2] (though whether they legally can require that is disputed [3]).

Content is owned by the user posting it, not Stack Overflow.  It is unclear whether asking a user to dual-license content is considered appropriate behaviour.

Small snippets may not be copyrightable, but it is unclear what counts as "small" in this context; a 9 line function has been a subject of litigation [4].  There have been multiple discussions of this on Stack Overflow meta [5].  I am in England, which has been said to have an unusually low threshold for copyrightability [6].

As a packager, how do I decide which ones are long enough to require action, and of what kind? - Remove them and write a replacement?  Risks breaking things (e.g. because I'm not sure exactly what the snippet is supposed to do) and/or still being similar enough to be an illegal copy (because I had to look closely at it to determine that). - Mark the package "License: (upstream license) and maybe CC-BY-SA" (is there a better syntax for that?) and report them upstream without a patch?  Also add the full [2]-form attribution (which might actually take more time than rewriting)?

[0] codesearch says 10988 occurrences of "stackoverflow.com" in 3036 packages.  Examples:
[1] https://meta.stackexchange.com/questions/344491/an-update-on-creative-commons-licensing
[2] https://stackoverflow.blog/2009/06/25/attribution-required/
[3] https://meta.stackexchange.com/questions/209250/stack-exchange-should-stop-using-the-creative-commons-logo [4] https://arstechnica.com/tech-policy/2012/05/oracle-to-pursue-longshot-claim-for-copyright-damages/
[5] e.g.
https://meta.stackexchange.com/questions/295599/short-code-snippets-licensing-on-stack-overflow https://meta.stackexchange.com/questions/12527/do-i-have-to-worry-about-copyright-issues-for-code-posted-on-stack-overflow https://meta.stackoverflow.com/questions/286582/can-we-get-some-explicit-clarification-on-the-intended-legal-usage-of-code-fro [6] https://commons.wikimedia.org/wiki/Commons:Copyright_rules_by_territory/United_Kingdom#Threshold_of_originality

The question is not how much you can use; the question is what you do with it.   How much you copy is only evidence to prove to a court whether something was copied in fact or whether is was created independently and is merely similar.  The longer the identical parts, the more probably that the work was copied.  That is a proof problem not a law problem.  The law problem is whether a work is copyrightable at all, whether it has been abandoned to the public domain and whether the use made of the piece is a Fair Use.  Length is not a requirement.


By the way,

Vincent F. Heuser, Jr.
Hirsh and Heuser Attorneys
3600 Goldsmith Lane
Louisville,  KY 40220
(502) 458-5879

Reply to: