Kirk Scott

Resources for instructors of capstone courses in computing

Most computing programs now have some form of integrative or capstone course in which students undertake a significant project under supervision. There are many different models for such courses and conducting these courses is a complex task. This report is intended to assist instructors of capstone courses, particularly those new to the model of teaching and learning inherent in the capstone course.This paper discusses important issues that must be addressed when conducting capstone courses. These issues are addressed through a series of questions, with answers reflecting the way that different institutions have chosen to handle them, and commentary on the impact of these different choices. These questions include: Goals of the Course; Characteristics of Projects; Project Deliverables; Sponsors; Teams; Prerequisites and Preparation; Grading and Assessment; Administration and Supervision; and Reflection, Analysis and Review.Subsequently we present information about the companion Web site, intended as an active repository of best practice for instructors of capstone projects. The Web site will have examples of information about capstone courses and materials used by instructors. Readers are invited to contribute content to this site. The paper concludes with a bibliography of additional reference material and resources.

DVH: a query processing method using domain vectors and hashing

The authors introduce a fast, space-efficient technique for accelerating equijoins between very large relations. The technique, called Domain Vector Hash (DVH) join Acceleration, is described and compared with three other join accelerators: hybrid-hash, join indexes, and materialized views. An analytic cost model is developed for each join method, and a detailed analytic performance comparison is made between each technique. The results show that DVH-join produces much faster joins than any of the other techniques considered, and over a considerably wider range of join selectivities and memory sizes. Moreover, the approach achieves this performance improvement while simultaneously minimizing the amount of additional join information that is cached on disk. When compared with either join indexes or materialized views, DVH typically requires from one to four orders of magnitude less cache space.<<ETX>>

Teaching graphical interface programming in Java with the game of wari

This poster gives an overview of a programming project that can be used in a course on object-orientation and graphical interface programming in Java.

PINE: Podium Incremental Neighbor Evaluator for classifying spatial data

Given a set of training data, nearest neighbor classification predicts the class value for an unknown tuple X by searching the training set for the k nearest neighbors to X and then classifying X according to the most frequent class among the k neighbors. Each of the k nearest neighbors casts an equal vote for the class of X. In this paper, we propose a new algorithm, Podium Incremental Neighbor Evaluator (PINE), in which nearest neighbors are weighted for voting. A metric called HOBBit is used as the distance metric, and a data structure, the P-tree, is used for efficient implementation of the PINE algorithm on spatial data. Our experiments show that by using a Gaussian podium function, PINE outperforms the k-nearest neighbor (KNN) method in terms of classification accuracy for spatial data. In addition, in the PINE algorithm, all the instances are potential neighbors so that the value of k need not be pre-specified as in KNN methods. By assigning high weights to the nearest neighbors and low (even zero) weights to other neighbors, high classification accuracy can be achieved.

Methods for distributed join processing using a voice-data protocol

The authors consider the problem of optimizing join query processing in a database distributed over a bus type local area network which uses the carrier sense multiple access/collision detection (CSMA/CD) access protocol. Some new algorithms are proposed which use a compatible access protocol, movable slot time division multiplexing (MSTDM), to achieve improved performance over existing algorithms. Analysis of example cases shows the improved performance potential for MSTDM. It is concluded that the proposed algorithms explicitly account for packetization and other costs unaccounted for in existing algorithms. If the overhead of CSMA/CD and MSTDM algorithms is comparable, MSTDMs performance characteristics translate directly into improved distributed join processing.<<ETX>>