Brian Hurt <bhurt@spnz.org> writes:

> On Wed, 20 Oct 2004, David Brown wrote:
> 
> > On Wed, Oct 20, 2004 at 06:30:01PM +0100, Richard Jones wrote:
> > 
> > > Just my personal opinion, but I've seen a lot of terrible code written
> > > which uses "design patterns" ...  A lot of the patterns seem to exist
> > > solely to bypass problems with OO languages.
> > 
> > Design Patterns seem very similar to what is usually called idioms in other
> > languages.  They're just common constructs you learn for common tasks.  An
> > example in ocaml might be typical recursion for a loop.
> > 
> > However, the OO Design Patterns do seem more to be techniques used to
> > compensate for defects in the OO methodology, or specific languages.
> > 
> > I would suspect that when someone encounters a place to use a design
> > pattern in OCaml, the problem would much better be solved by using some
> > other feature of the language.
> > 
> 
> I don't seem to have quite as negative a view of the GoF book as many 
> people around here seem to.  But I think this is because I see design 
> patterns as a more general idea than just what the GoF presented- and I'd 
> argue this view is supported by the pattern community- as now you have 
> software architecture patterns, anti-patterns, etc.  I think there are OO 
> patterns (which is what is documented in the GoF book), functional 
> patterns, even procedural patterns.

I totally agree with you.
In fact the haskell community have developped design patterns at:
  http://www.haskell.org/hawiki/CommonHaskellIdioms

They have adapted the idea of GoF to push it further cos in functionnal langage the main
ideas of the GoF book were already known/used by the fp community.



> 
> I also think that the difference between idioms and patterns is mainly a
> matter of degree- they're all of a kind, if you see.
> 
> Note that there are a number of Functional patterns which are for working 
> around the limitations of the Ocaml language as well.  An example I'll 
> hold up here is "build a list backwards, then reverse it" pattern, to make 
> a function tail recursive.  Others are less clear cut- is "return the 
> update applicative structure" a work-around, or a better way of doing 
> things?
> 
> The point I'd like to make to the original poster is that the GoF isn't 
> the last word on how to think about problems, especially not in Ocaml.  
> Look for the new patterns Ocaml allows/encourages.

People generally disrespect the work done by other communities (and then
reinvent it giving it another name).

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners