On 12/9/23 13:30, Tim Parenti wrote:
It still falls short, though. of the ideal that the leap-seconds.list file itself also contain the notice so we (and others) don't have to duplicate it elsewhere: https://mm.icann.org/pipermail/tz/2017-April/024988.html
It also doesn't necessarily allay the other concerns.
Perhaps we could work around the copyright-notice problem by adding appropriate verbiage to the LICENSE file. This would require some thinking, though. One other problem just occurred to me. Since the IERS variant of leap-seconds.list is not public domain, putting it into tzdata would contradict Internet RFC 6557 section 7 "Data Ownership"[1], which says that tzdata is public domain and therefore is exempt from a bunch of rules that I'm not familiar with and would rather not become expert in if I can avoid it. I suppose this could be worked around by writing a new RFC that addresses this issue but that'd surely be a good bit of work in its own right. For now, I expect we're better off sticking with the NIST variant since it's public domain. If NIST stops distributing their variant then it might even be better for us to distribute just the *data* (which are identical in the two variants - they differ only in comments). We could do that by removing comments and recalculating the checksum as needed. Although I wouldn't like this solution, that's often the case when lawyers are involved.... [1]: https://datatracker.ietf.org/doc/html/rfc6557#section-7