Re: Interview question

On 30-Jun-2016 09:27 -0500, John Yeung wrote:

<<SNIP>> I suppose the verbiage of the instructions contains
cultural, educational, and linguistic biases, the same way many,
many standardized test questions do. <<SNIP>>

I suspect that their choice of the term /delimited by/ is one such example. I eventually concluded that, most likely, they meant to have used the term /separated by/.

IME the computing vernacular has defined the former term as fully enclosing the boundaries [of the string of alpha characters for this scenario], such that both beginning and end must have the non-alpha; their example, "Automotive"->"A6e" seemed to emphasize that intention. I consider the latter term to imply only one boundary of the string of alpha characters, either the beginning or the end, need be demarcated by a non-alpha; e.g. with comma _separated_ values, wherein some of the values additionally may be _delimited_ to avoid a reader mistaking a separator-character that is embedded within the delimited-value, as an actual separator-character denoting the next value.

Out of curiosity, which definition did you use, for the solution you started coding? Did you ever imagine any alternatives to your first choice? I would guess that if your first inference was /separated by/, then no other alternative definitions were considered.? If the first interpretation of /delimited by/ is a requirement to enclose, then the programming task has a number of error cases to which no allusions were made in the rules\description of the problem domain.