An interview with Larry Wall (LinuxVoice)

The number of different ways to do something provided by a language also depends on what is included in 'the language'. The C language itself does not provide data structures such as linked lists or hash tables, or any input/output primitives. Once you add the standard C library you get the question of whether to use puts() or printf() - or sprintf() followed by puts(), and so on. Add a library for data structures and you then can choose whether to use a vector, a linked list, a tree... These combinations multiply together pretty fast to give easily 'dozens' of ways to achieve a fairly simple operation such as picking the top three integers from a sequence. With a higher-level language, more data structures and manipulation operations are part of the core language and the syntax tends to be more compact, so it is not a surprise to see many different ways to express even a single line of code.

Take Python as an example, often held up as a contrast to Perl. "There should be one obvious way to do it" is certainly a fine principle and one that Perl could sometimes take note of. But in Python too you can choose between lambdas, list comprehensions, explicit loops (of several types, with some interesting ways to break out of the loop); lists, dictionaries or perhaps tuples; and of course different ways to structure object-oriented code. Again there are easily dozens of ways to achieve a relatively simple operation. And this is a good thing.

I don't have experience with Perl 6 so I can't comment on details, but I will say that if the language includes more powerful tools like grammars, these obviously provide new ways to tackle a problem, and so several more "ways to do it".