(1) make-bitvector's fill argument is a bit (i.e. 0, 1, or a boolean),
as are the arguments of `bitvector'. I suggest the names "bit"/"bits"
instead of "fill"/"values", respectively.
Fixed.
(2) It seems that bitvector-reverse! and bitvector-field-reverse! do
the same thing, except that the start and end arguments of
-field-reverse! are mandatory. Are both necessary?
Removed the -field- versions.
(3) Just as with bitvector-unfold, we should unify the unfold-right/{int,
bool} procedures and provide a single right unfold. This has already
been done in the sample implementation.
Done.
(4) bitvector->vector/{int, bool} and vector->bitvector should have
optional start and end arguments.
Added.
(5) The meaning of the optional len argument of bitvector->integer
should be documented.
I have no idea. I think it reflects a confusion on my part: it should have been an optional argument on integer->bitvector, so that you can specify the length of the bitvector you want to have (but it is still bounded by integer-length by default). I have removed it from bitvector->integer and added it to integer->bitvector. Let me know if you disagree.
(6) I believe John decided that the bitvector-generator functions
should be renamed to make-bitvector/{int, bool}-generator. I've
made this change in the implementation, but not it seems not to
have made it into the document.
Fixed, and all changes pushed.
John Cowan
http://vrici.lojban.org/~cowan xxxxxx@ccil.orgThor Heyerdahl recounts his attempt to prove Rudyard Kipling's theory
that the mongoose first came to India on a raft from Polynesia.
--blurb for Rikki-Kon-Tiki-Tavi