change: description and examples do not match

[abeger]

The description states that the output is an array that lists the amounts of each coin needed:

• An input of 15 with [1, 5, 10, 25, 100] should return one nickel (5)
  and one dime (10) or [0, 1, 1, 0, 0]
• An input of 40 with [1, 5, 10, 25, 100] should return one nickel (5)
  and one dime (10) and one quarter (25) or [0, 1, 1, 1, 0]

However the tests state that the output is an array of individual coins

      "input": {
        "coins": [2, 5, 10, 20, 50],
        "target": 21
      },
      "expected": [2, 2, 2, 5, 10]

[petertseng]

Anyone who has opinions on which it should be should speak now. It is known that the tracks that had this exercise in July (the most recent data available) were all using list of coins, such as [5, 10], rather than lists of counts, such as [0, 1, 1, 0, 0].

the list of tracks that had this exercise in July:

clojure/exercises/change
csharp/exercises/change
ecmascript/exercises/change
elixir/exercises/change
fsharp/exercises/change
go/exercises/change
haskell/exercises/change
java/exercises/change
javascript/exercises/change
kotlin/exercises/change
lua/exercises/change
ocaml/exercises/change
php/exercises/change
python/exercises/change
reasonml/exercises/change
ruby/exercises/change
scala/exercises/change

[petertseng]

I would consider a list of coins clearer than an uncontextualised list of counts, so I prefer the list of coins.

Another option I would consider is a mapping of coin to count. I'd accept {5 => 3, 10 => 4} to represent [5, 5, 5, 10, 10, 10, 10].

[ErikSchierboom]

I also prefer the list of coins, as it best represents one having actual change in their hands I think.