On 6/3/19 12:11 PM, Enrico Weigelt, metux IT consult wrote: > By the way: does it still make sense to keep all the "copyright(c) ..." > lines, or would should that information always be retrieved from git ? > (for that particular file, git blame doesn't show anybody who seems to > be from broadcom) Yes, I think they would be. That being said, copyright notices in the Linux kernel are horribly out of date/inaccurate and even Git does not do a good job of recording these (garbage in, garbage out). Git only has a single "Author" field and I know of commits where the commit message had two authors listed and the that was added had three authors listed. Then there are all kinds of silly tags in the Linux kernel where it is absolutely unclear whether or not someone is claiming copyright or not. See for examples this article: https://www.ifosslr.org/index.php/ifosslr/article/view/121 Also, author doesn't necessarily mean copyright owner, company affiliation also does not necessarily mean copyright is with the company (it depends on the work contract) and it would require a lot more research to find the true copyright owner. The copyright statements also have other relevant information, such as years. This is relevant to find out which copyright laws were in effect when the code was published. > Or should we introduce an SPDX header for that ? There was a discussion about this a few days ago (please see the list archive). Personally I don't see the benefit. armijn -- Armijn Hemel, MSc Tjaldur Software Governance Solutions
