John, On Tuesday 19 April 2005 00:10, John Lamb wrote:
Colin Carter wrote:
This is a good site John, I saw a good programme on the BBC about this. It disclosed a few secretes - I wish I had taped it. I couldn't see anything on this site about the 'bug', but I think that the BBC programme mentioned it. Any ideas?
I'm sure I saw the same programme but I don't recall the 'bug'. I guess a Google search might throw up something. There are also several books, which are bound to go into more detail. IIRC the programme was called Station X and is the sort of thing that gets repeated regularly on various history channels.
I believe object-oriented programming developed from an algol based language called Simula in the 1960s. Simula was designed for discrete event dimulation
GSS is not too bad for a quick solution.
I don't recall what GSS is. For fast discrete-event simulation I now usually use C++, though Java is also good. The only thing you need that isn't pure O-O is a priority queue.
I really wish people would stop using the phrase "pure Object-Oriented." The word "pure" should only be used in the context of chemistry, if you ask me. Applying it to programming or, god forbid, ethnicity just invites trouble, or, at best, confusion. Anyway, the Concurrent Programming in Java class library (now incorporated into Java 1.5 / Tiger) has a perfectly serviceable heap-based priority queue ("heap" as in the data structure, not the dynamic storage allocation mechanism). I use it extensively in my theorem prover and apart from exhibiting non-stable ordering, it works great and performs well. I created a modified version that is not synchronized so I could avoid the overhead produced by synchronization (though Sun claims that late-model JVMs have near-zero synchronization overhead). I sometimes move hundreds of thousands of items through these queues in a single proof and they do not become a hot-spot in that application.
...
-- JDL
Randall Schulz