Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!pdn!tscs!tct!chip From: chip@tct.com (Chip Salzenberg) Newsgroups: comp.lang.misc Subject: Re: CHALLENGE: heterogeneous collections Message-ID: <27F77BD2.1741@tct.com> Date: 1 Apr 91 18:28:34 GMT References: <27F11E1D.6354@tct.uucp> <49413@nigel.ee.udel.edu> Organization: Teltronics/TCT, Sarasota, FL Lines: 32 According to carroll@cis.udel.edu (Mark Carroll): >The argument shouldn't be: > "I can do heterogeneous lists in a static system (using MI), so > since you don't need dynamic typing to do it, dynamic typing is > no good". > >The correct form of the argument is this: > "My staticly typed solution is a BETTER solution than your > dynamic one, and this is why:". I cannot cast my argument in this form, because that's not the point of this thread. The original question was, "Why do you _need_ heterogenous collection?" I have enjoyed debunking some of the replies. However, the heterogenous/homogenous collection issue is really a side issue for me. I like static typing for its measure of safety from message-not-recognized errors. If that forces me to use the C++ preprocessor to make collection classes, I can live with that. >Now, in the context of this argument, I think that the dynamic solution >under which I can iterate ANY method over a heterogeneous lists is a >better, cleaner solution that the static solution in which I have to >create an artificial superclass for every method that I want to use >to iterate over the list. Agreed, as far as it goes. But I'm not willing to sacrifice type safety for that notational convenience. -- Chip Salzenberg at Teltronics/TCT , "All this is conjecture of course, since I *only* post in the nude. Nothing comes between me and my t.b. Nothing." -- Bill Coderre