Problems and solutions?
Recently i came across a couple of items that stuck in my mind initially because they had ‘funny’ names. One was ‘Byzantine Fault Tolerance’ and the second was ‘The German Tank Problem’. Also thrown into this mix was a set of problem spaces i read about called ‘Fermi Problems’. This interest also stems form a longer interest in problem solving by using structured processes or mathematical and statistical processes to deal with the problems.
For instance the whole field of Operational Research which largely came into being in World War Two to help solve the problems of convoys and the large movements of troops and materials required for operations such as Overlord and the USA’s conflict in the Pacific with Japan where distances were so large that some method of rigorous and provable planning was required. Of course this then leads onto the development of computers both mechanical and electronic.
The idea behind the site is not to repeat large amounts of information from other sites which certainly do a better job of explaining in detail the methodology but to enumerate the different types and list the various resources available and the general class of problem they solve and i freely admit this is mainly for my own benefit as i intend to produce iPhone/Pad apps that deal with some of these solutions.
Why a blog format? Well, trying to structure this into a web site would give me a lot of pain and take more time that the structure would be worth. So the process of doing it in a Jekyll static blog appealed and i can cross reference it as much as i like as i go to give me a beet cohesion.
To put this whole thing into a bit of perspective i will quickly go over the three items that sparked this blog.
This is a class of problem for which there is little or no data with which to work with. So to solve a Fermi problems you need to make a series of smaller steps, each of which you can more or less justify until you solve the problem. In Wikipedia,
German Tank Problems
Byzantine fault tolerance
Coming a bit later