I tend to agree that 'pad' implies adding zero or more elements but does not remove any. In fact Gauche has implemented so, but a user pointed out ".. so that the length of the result is _length_" text, then I found that the string-pad does truncate.

If both of the authors like to make it an error when _length_ is shorter than the bitvector, it's ok as well, though inconsistencies between SRFIs may confuse newcomers. The SRFI text should be clarified either way, though.

BTW, the current reference implementation returns _bvec_ as is when _length_ is the same as the length of _bvec_, but it contradicts the SRFI text "Returns a copy of _bvec_ ...". As bitvectors are mutable, this makes a difference.

On Mon, Aug 28, 2023 at 9:11 AM Wolfgang Corcoran-Mathe <xxxxxx@sigwinch.xyz> wrote:

Sorry. I realize that Shiro already posed the same question.

While string-pad is precedent, I wonder if "pad" is the right name
for a procedure that may truncate a bitvector. "Padding" suggests
to me that the result will always be at least as long as the input.
Unless the name is changed, I lean toward returning the input vector
or making it an error if *length* is less than input length.

--
Wolfgang Corcoran-Mathe <xxxxxx@sigwinch.xyz>