In Software Development (U.S. Software Engineering), there are basically three different models of the process of development:

Waterfall model, Evolutionary and component-based development.

Waterfall model identifies different phases which are reviewed in all development projects. This consists of a requirements specification, design, implementation and unit testing, integration and acceptance testing, and development. All the steps are reviewed sequentially and deliverables in the form of reports presented at each stage is completed. There is thus no iteration between the phases and the management can easily control the process. This has obvious advantages but also disadvantages in a changing environment.

The second conceptual model of development is evolutionary development. It consists of the same phases as the waterfall model, except that these are reviewed in several increments. After each version is the customer and creating new demands for the next version which is then designed, implemented and validated. The model can be used to develop complete systems as well as prototypes for requirements specification. Although variants of evolutionary development is located. Examples are iterative development where the requirements are completed and the phases are then examined in several steps of iterations. Another example is the spiral model where a lot of freedom.

The third main software development model is component-based development (CBSE). The development branch is to be used the final so-called COTS products, commercial of the shelf. The process has been going through the following stages: requirement specification, component search, claims adjustment to the items found, component selection, integration of components and the validation of requirements. Using CBSE minimize lead times and costs, reduce risks of project failure and increases reliability because the component is tested before. Disadvantages are that the code usually can not be inspected because the component often comes in a compiled format. This results in poor control of maintenance and what will happen in future versions. In addition to the components are executable (compiled) is the standard, whatever else, interface (requires and Provides) and are well documented.

I work a part of java and the summer working with matlab. Using a waterfall model, then, because of the workload is quite impossible. This summer, I have a month on me to put together a matlab program to be graphic to be used for calculations of various eclectic system. Doing this from square one would have required time to learn how everything is in place, build capabilities, requirements, architecture, etc.. I will assume functions that already exist and my job is to make the interface to enter these functions so that they can be used in a better way. Component-based development ..

GPS clock

I've been thinking about various private projects. I started a bit of a flash application in order to present cool graphs from data from a database, but quickly realized that it did not interest me. I've been thinking about creating my own program for GPS processing. What I use now is focused on training and has some good features, but I think I could create something much better. The idea is to write program in java to use it both locally and via an applet on the web. The program will combine regular gps management in the forest environment and training with GPS like my GPS watch (←).

I think I have mastered most of the techniques will be needed to create the application. I have nylighen found a very good open component for GPS maps that I will use. The component encloses (wrapper) javascript, which is used for programming for most maps to exepel google maps , with java. Claudius Hauptmann is another good resource written a few articles about programming to google maps.