Mike Meyer wrote:
Saving time for people who are used to having a well-designed hierarchy of collections by letting them not waste time dealing with name differences over implementation details. Things like repeating myself by writing foo_for_sets and foo_for_dicts and foo_for_lists, or changing method names, or mapping between two representations so I can use code that chose one set of names, or dealing with the headaches of maintaining two copies of a structure.
Blurring semantic distinctions (add != append) to make overgeneralisation easier doesn't sound like a good idea to me. Choose the data structure that makes the most sense for the current algorithm. If someone gives you a different data structure, convert it at the beginning and convert it back at the end (think of it as a variant of the decorate-sort-undecorate approach to sorting). Cheers, Nick. -- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia ---------------------------------------------------------------