this means the only way to trigger the behaviour on purpose is to pass undef as first parameter. API madness! [...] this behaviour of implicit $_ is not documented at all.

How can something that's not in the API (undocumented) be the basis of a claim of API madness?

I cannot imagine how dismantle is useful.

my @elements = split //, $s;
my $length = @elements;
[download]

could be written as

my ($length, @elements) = dismantle($s);
[download]

but then again, you can already write

my $length = my @elements = split //, $s;
[download]

or

my $length = my @elements = $s =~ /./sg;
[download]

since list assignment in scalar context returns the number of elements assigned.

---

Problems you didn't mention:

• trim and strim don't work when passed an empty string.
• trim and strim don't work when passed the string '0'.
• strim doesn't work if the $delim is more than one character long.
• strim could be written much better.

Problems you mentioned I agree with:

• Allowing undef for the first parameter is unclear, wordy and error-prone.
• Warnings usually provided by substr are not emitted.
• The trim function is badly named.
• The module is poorly named.
• strim doesn't properly convert the delimiter from text to a regex pattern.
• trim could be written much better.
• trim is exported by default.
• strim and dismantle aren't exported.
• The synopsis does does not conform to best practices.
• Badly formatted docs.
• Functions are of very limited usefulness. It's fine to write a module that has easy to implement functions, but not when they are harder to learn than the building blocks to build the functions in the first place.