Hello again,
I think that I have done all of the above in different situations, and so I don't think that there is a single correct answer. I usually avoid using the "newtype" trick as I find it inconvenient: usually the newtype does not have the same operations as the underlying type and so it cannot be used directly, and if you are going to wrap thing just when you use the class methods, then you may as well use the non-overloaded operations. Yes, I had the same thought. If I have to explicitly convert between various wrappers, that's basically the same as defining separate functions from the start. The newtype trick is certainly nice when you don't want to mix the different uses on the same piece of data, but that was exactly my problem. I'd say your posts as well as wren ng thornton's answer my question as thoroughly as possible at this point, although wren's lengthy letter could probably even be extended and turned into a tutorial on the topic.
Of course type classes are not only important just for making dictionary passing implicit, but also because of their interactions, which is nicely demonstrated by the sigfpe blog entry that someone also linked from this thread. That's probably why I feel kind of uncomfortable when not taking advantage of standard classes wherever possible. Thanks for the answers, Gergely -- http://www.fastmail.fm - Access all of your messages and folders wherever you are