Abstract

With vast amounts of data in the world, organization becomes a challenge. The success of data driven web services (IMDb, YouTube, Google Maps, Wikipedia, et cetera) all hinge on their ability to present information in an intuitive manner with user friendly interfaces. One area that fails to have such a service is sports statistics. With the ubiquitous appeal of sports, having a solution to this problem can be universally beneficial. Many sites exist that have statistics of different sports, but there are limitations to all of them. Since there is very little continuity among all sports, statistics are represented disparately.

There are several problems with this approach. Any time there needs to be a change to the informational structure, the entire database and interface need to change. In addition, there can never be a single interface if there are different schemas for different sports, leading to a user unfriendly interface.

My system uses a unique schema that is capable of representing statistics from any sport, no matter how unique. Adding new statistics to a sport to reflect rule changes or adding a new sport altogether are seamless. In addition, the web interface is structured by Rails, which changes automatically with the schema.

Challenges included developing a universal sports schema and testing it sufficiently enough to prove its generality. Finding and extracting the data to populate the database also presented difficulties.

Abstract
 
This thesis examines and investigates the substitution of the mouse for a more natural means of human computer interaction, ranging from manipulating WIMP-based applications to developing post-WIMP interfaces and exploring their usefulness. The WIMP (Window, Icon, Menu, Pointer) interface has been the standard paradigm for the personal computing era. Computer software has been optimized for the keyboard and mouse input device pair, tools that have remained fundamentally stagnant with regard to innovation for decades[1]. Accomplished through the construction of a touchscreen with variable levels of contact detection, targeted demo applications not only show the effectiveness of such an input apparatus but introduce the potential for previously unexplored levels of interaction.

The use of direct-contact manipulation provides a more natural interaction than is achieved by the mouse, avoiding the need to abstract crucial concepts such as 'selecting', 'dragging', or 'resizing'. The introduction of vision driven touch-sensitivity allows for the use of physical objects to denote distinct meanings, providing a means to create associations between physical and digital actions. Building upon this concept, gesture support is a logical and practical capability to expect from a 'direct' input device. As such, it is analyzed and implemented as a core component of the device's software.

Common difficulties with software based touchscreens include device mobility, device reliability, and poor interface software implementation. While mobility is not mitigated within this project, reliability and interface/usability design are principally addressed. Challenges addressed during the implementation of the project primarily revolved around physical limitations and performance restrictions, as the quality of algorithm necessary is inversely proportional to the quality of equipment being used.