This situation is all over the undergrad academia. For a math major, numerical method algorithms described in advanced notation is ok, as long as you are fluent in it. But when I take CS courses and see the same presentation, I get very confused. For god's sake, these people are going to spend more time on Novell manuals, than on the graph theory. I could have learnt all these algorithms, if I didn't need to chew through all that theoretical packaging and generalization. Instead I spend all my time to dig through precise definitions just to get one or two concepts, when I could have gotten them all, if representation was good enough.
I don't think that CS undergrads don't need precision offered by scientific approach. I think there comes a time when you feel the need for precision. But this time won't come when you only know quicksort in java. You have to gain a broad specter of examples before you can set off to learn about big O analysis. Because this kind of analysis is about comparison. Learning is ineffective when you compare one algorithm with a blurred idea of another.
I might sound like a confused student, but what I'm trying to say is that there should be one unified, simple and consistent source of algorithms in understandable languages. I have the large algorithm encyclopedia from Springer Verlag, and it polluted with all uncertainties of the modern CS studies. It shows us the perspective fields of study, but it's too abundant to explain simple ideas. Same thing is about Wikipedia. It tries to be both precise (because experts write articles) and simple (because people like me read them) - but this is like running in opposite directions simultaneously.