What is a good delimiter for Perl?

That's not an interesting question. You shouldn't choose your delimiter based on the language the program(s) that process the data, but on the data itself.

When I have to pick a delimiter, the questions I ask are:

1. Do I want to look at the data myself?
2. Is the data generated or edited by humans?
3. Are there characters (or short sequences of characters) that cannot occur in the data?
4. If the answer of the previous question is 'no', what delimiter is expected to be rare? (I also have to start thinking about an escape sequence).
5. Do I want the delimiter to be surrounded by optional whitespace? (If any of the first two questions is 'yes', the answer to this one tends towards 'yes'; if the data can have fields with leading or trailing whitespace, the answer tends towards 'no').