Am Do., 20. Aug. 2020 um 17:25 Uhr schrieb John Cowan <xxxxxx@ccil.org>:

>> > Three, I think: reverse-bitvector->vector/{bool,int} and reverse-vector->bitvector.
>> s/vector/bytevector/g   ???
>
>
> I am not sure I understand.  We already have a bitvector-reverse, as well as procedures to convert to and from lists both forward and reverse.  We only need to extend that capability to vectors.

Ah, I think we may have talked past each other. While I actually
didn't notice that the procedures converting to and from vectors have
no "reverse" equivalent, I have been concerned about the
"bitvector<->bytevector" procedures, which currently only use
big-endian order according to the document. When talking to other
clients (FFI, over the wire, files, whatever), encoding bitvectors as
bytevectors is probably one of the most common things, so I wanted to
make sure that fast routines are available for the most common formats
of encoding bitvectors as bytevectors.

> Institutional persistence, probably, as well as institutional OCD.

:-)

> The time to do so came in the U.S. on January 1, 2019, and no such revision was introduced.  So everything first published in the U.S. before 1925 is in the public domain, after many years of the limit being frozen at 1923.  (In particular, the very earliest versions of Mickey Mouse, an incredibly valuable property, are P.D.)  Congress can and has, however, restored copyright status to works before, but I know of no efforts in that direction.

I had the graphic on this Wikipedia entry in mind:
https://en.wikipedia.org/wiki/Copyright_law_of_the_United_States#Duration_of_copyright.

> (Unpublished works can remain in copyright forever.  There exists a letter by Thomas Jefferson (1743-1826), written before there was a United States, that will remain in copyright until 2048, as the original Province of Massachusetts Bay perpetual copyright on unpublished works was eventually transferred to a historical society, which published it in 1958 -- on microfilm only.)

Strange.