svn commit: r314905 - in head/sys: compat/linuxkpi/common/include/linux compat/linuxkpi/common/src conf modules/linuxkpi

[Rodney W. Grimes]

> On Wed, Mar 8, 2017 at 10:05 AM, Warner Losh <imp at bsdimp.com> wrote:
> >>> The project's guidance to
> >>> committers for the last 20 years is to do a range of copyright dates.
> >> The projects guidance has wrongly been changed then, as I have
> >> always tried to make sure the A, B, C-D information was applied
> >> correctly.  I do have a fairly good understanding of copyright law.
> >>
> >> Can you point to any "published" project guidance on this manner?
> 
> When the guidance was originated, we used
> https://www.oppedahl.com/copyrights/ as a guide since I knew Mr
> Oppedahl personally and we talked about it at the time. He recommended
> that we use ranges, as he does in his FAQ. He said it was the safest
> way to not mislead about the copyright dates. Microsoft uses ranges of

His FAQ does agree with me that the date is one of the most important
parts of a copyright notice, and that it can easily be a pitfall:

(Sorry for not wrapping this)
The date. The copyright date is perhaps the most important trap for the unwary. One of the purposes of the copyright date, under U.S. copyright law, is to assist members of the public in identifying works which are so old that the copyrights have expired. To do this, a member of the public would take the copyright date appearing in the notice, add to it the number of years of the copyright term, and thereby arrive at a conclusion as to when the copyright would have expired. In the case of computer software, it is common place for the work to include original matter from many different dates including original work dating from any of several different years. Consider what would happen if the most recent year were the only year used in the notice. A member of the public would then be led to the conclusion that the entirety of the work is protected by copyright starting from that year and ending at the end of copyright term. But if part of the work dates from a previous year, then its term expires one year earlier than the rest of the work. This could mislead members of the public in the sense that they would incorrectly think that none of the work could be copied until the end of the term that is based on the date in the notice, when in fact part of the work would have entered the public domain one year earlier than the end of that term. There have been court cases where judges have stricken all of the copyright rights in a work due to such incorrect statements in the copyright notice. 

His FAQ does not actually recommend ranges, it only states this is
another thing done:

Another approach is to put a range of years. For example, if the oldest matter in the work dates from 1991 and if the newest matter dates from 1994, the notice might say copyright 1991 to 1994 and the name of the owner.

I assert again the problem with a range is that material inside
the range may actually be reaching the end of the time a copyright
is valid, and to have the public misslead about that is... well..
wrong, and can and has lead to legal consequences.

Quuoting your Mr Oppedahl (also appears above ):
There have been court cases where judges have stricken all of the
copyright rights in a work due to such incorrect statements in the
copyright notice.

> dates, even when they haven't made changes in every single year. Of
> course, talking to a lawyer about this gives one a big "it depends"
> and things get fuzzy. The practical implication might be an inability
> to enforce the license terms at the end of the 90 years that people
> have copyrights for, so as a practical matter he suggested that for
> open source a range was the best compromise between an exhaustive list
> of years and never updating the notice.
> 
> It is (or at I think it was) in the developers portion of the
> handbook, but I can't find it now. It's implicit in style(9) (I made
> the changes there). It's come up several times in the past.
> 
> But like I said, please feel free to improve things by getting a
> definitive statement that our current range is wrong from the lawyers
> and getting their recommended advice. However, this has been SOP for
> the last 20 years, so there's many places that would need to be
> corrected.... I doubt that's it's a windmill worth tilting at, but
> it's your time.

I believe there is a 5 year period that these could of been corrected,
I am not concerned about that, cant do much about things loss due
to poor SOP, I can, and well, tilt this windmill.  

> Warner
-- 
Rod Grimes                                                 rgrimes at freebsd.org