On Thursday 04 December 2008 21:46:37 Andreas Hauser wrote:
jimka wrote @ Thu, 4 Dec 2008 14:43:35 +0100:
* A very opinionated and probably objectionable view of CLOS (the way I think it should be).
A CLOS discussion would be nice to have. I remember some other people mentioning they've never come around to dive into it. I'm a bit uncertain about your stance (some native speaker connotaions might have been lost), are you pro or con CLOS? Can someone else defend the other position?
Yes, i'm basically Pro-clos. However, i do no know a lot about theoretically sound object models, and cannot answer questions about CLOS in that regard. I look at CLOS as a tool box of lots of capabilities that just very often come in extremely handy. There are, however, some design decisions which went into its original design (usually for performance reasons) which I think are unfortunate and others which I think are obtuse. One thing that I think is interesting about CLOS is its ability to do what I call Specialization Oriented Programming. I.e., using CLOS not for its object model (where classes represent state laden objects with similar characteristics), but rather using the method dispatch protocol for control flow purposes. In this approach objects of a class do not represent entities in any describable sense, but rather they act as policies to control which methods are applicable and in which order they are evaluated. I suspect many people will disagree with my opinion. I look forward to the discussion -jim