[Biojava-l] Tutorial Suggestion...

[Phillip Lord]

>>>>> "James" == James Carman <james@carmanconsulting.com> writes:

  James> No, each flyweight instance is not a singleton.  It is a
  James> single instance of a class, that's true.  However, there are
  James> also other instances of the same class.  In the singleton
  James> design pattern, exactly one instance of the class exists.  In
  James> the flyweight pattern, multiple instances of the class exist.
  James> There are four AtomicSymbol instances created to represent
  James> each of the a, g, c, and t DNA symbols.

I think that the confusion comes from another related design pattern
which is a common way of faking enums in Java. So...


class Nucleotide{

        private Nucleotide(){};
        
        public static final Nucleotide a = new Nucleotide();
        public static final Nucleotide c = new Nucleotide();
        public static final Nucleotide g = new Nucleotide();
        public static final Nucleotide t = new Nucleotide();
}

is not a singleton either, but clearly looks like it. It's not really
a flyweight either, although it fulfils the same function where the
number of instances are known and defined at compile time. Personally
I consider it to be a singleton-ish method for nearly implementing a
flyweight.

See this is the problem with computer programming. Someone spends all
this effort on synthesising only 23 patterns out of enormous
complexity. And then someone at the back stands up and says "ah but
what about....".

A bit like biology really. There's always an exception to every rule. 


Phil