I’ve been interested in the core data structures and algorithms that pin applications together since shortly after I began university. I have always found it fascinating how a simple change of selection of data structure for a task can turn a painfully slow process into a blindingly fast one. Also how the same selection can change a task that was difficult to reason about into one that flows naturally. Around this is the design of how you interact with these structures and algorithms, which is equally crucial.

Over the past few years I have had a tendency to write my own libraries to do most general tasks, as often the mainstream (freely available) libraries are just a little inadequate (which gives me an excuse to have a bit of fun). I realised recently with the number of people beginning to use Google Collections, for instance, that people are finally realising the limitations of the JDK collections libraries, so I decided a little over a week ago I would redesign my own from the ground up again in order to produce something I think will be of value to Java programmers everywhere. I have chosen to call the resulting library “jjoost”, after the famous Boost libraries for C++.

This blog will initially be about why I think the existing frameworks are inadequate, how I think they can be done better, and the things I find interesting about the data structures I have been or am currently writing – or the technical challenges therein. Hopefully this will be of interest to some people!