"OOP is Dead" article
in [OOP]
|
From: topmind on 3 Feb 2007 22:38 http://kawagner.blogspot.com/2006/08/oop-is-dead.html The article complains about polymorphism's narrow dispatching capability and suggests functional programming has a better angle on that because it can use as many parameters as you want for dispatching (predicate dispatching). I had this to say about it in another blog: The problem with predicate dispatching is that in practice the implementations are not one-to-one with the various combinations of parameters (predicates). The implementations tend to interweave and are assymetrical such that one activity may involve many combinations and others only one. Saying "this block of code goes with combination A and this block goes with combi B" is not good enough. The granularity needed is still smaller than that. IF statements are still the best for that kind of thing in my opinion because you can have 10 or 1 per block, sub-block, sub-sub-block, etc. Sure, you could divide the code into micro- blocks, but it's a cure worse than the desease. Often the real world just does not offer nice chunk-able abstractions. IF statements deal with the slop fairly well. -T-
From: adaworks on 14 Feb 2007 02:55 "topmind" <topmind(a)technologist.com> wrote in message news:1170560281.725059.248210(a)l53g2000cwa.googlegroups.com... > > http://kawagner.blogspot.com/2006/08/oop-is-dead.html > > The article complains about polymorphism's narrow dispatching > capability and suggests functional programming has a better angle on > that because it can use as many parameters as you want for dispatching > (predicate dispatching). > Apparently someone has never heard of OCAML. Richard Riehle
From: topmind on 14 Feb 2007 11:57 On Feb 13, 11:55 pm, <adawo...(a)sbcglobal.net> wrote: > "topmind" <topm...(a)technologist.com> wrote in message > > news:1170560281.725059.248210(a)l53g2000cwa.googlegroups.com... > > >http://kawagner.blogspot.com/2006/08/oop-is-dead.html > > > The article complains about polymorphism's narrow dispatching > > capability and suggests functional programming has a better angle on > > that because it can use as many parameters as you want for dispatching > > (predicate dispatching). > > Apparently someone has never heard of OCAML. > > Richard Riehle Some suggest that such abilities are really functional or logical programming concepts put into an OOPL, and not really OO. But, let's not start another OO definition war :-) -T-
From: pandit on 20 Feb 2007 10:46 > Apparently someone has never heard of OCAML. hey Richard, i am also sad on this, "ocaml" is a very good language: http://caml.inria.fr/ unfortunately, most programmers never even heard of "Mercury": http://www.cs.mu.oz.au/research/mercury/
|
Pages: 1 Prev: Specification Pattern vs Data Mapper Next: Relational-to-OOP Tax |