Bookmark and Share
|
  • Text +
  • Text -

Computer Science

Programming in Grace

Amy Ruskin (2014); Student Collaborator(s): Richard Yannow (2014); Mentor(s): Kim Bruce

Abstract: Grace is a programming language that is currently in development with the eventual goal of being used to teach introductory computer science courses. I programmed extensively in Grace to find some remaining bugs and provide feedback on the experience of actually using the language. In the end, I produced Grace code for various data structures, using Java structures as guides, and translated some of the projects and assignments from Pomona’s Data Structures and Advanced Programming (CS62) into Grace. Most of the problems encountered were due to features of the language that were not yet fully implemented and the lack of extensive and current documentation, but once those issues are resolved, Grace should be easy to learn and straightforward to write.
Funding Provided by: Pomona College SURP

Adapting Object-Oriented Languages for Instructional IDEs

Richard Yannow (2014); Mentor(s): Kim Bruce

Abstract: The Grace programming language project was started with the intention of making a new object-oriented language for teaching the practice of programming. In order to be successful, Grace must be easily usable by novices, and a significant factor towards that goal is having a beginner-friendly integrated development environment (or IDE). We decided to use DrRacket as an IDE for Grace, allowing us to take advantage of its numerous novice-friendly features and Racket’s powerful language-building capabilities. We developed a new backend language, Racket-Grace, with Grace's semantics, but with Racket-style syntax. We also wrote a pretty-printer that will take processed abstract syntax trees of Grace code, and return an equivalent Racket-Grace program. This will allow us to input a Grace program into DrRacket and run it there, translating it into Racket-Grace as an under-the-hood intermediate step, allowing us to maintain compatibility with DrRacket's many useful tools.
Funding Provided by: Pomona College SURP

Examing graph clustering stability

Evan Fields (2013); Mentor(s): Tzu-Yi Chen

Abstract: A graph is a collection of objects (nodes) and connections between pairs of objects (edges). Graphs have been used to study phenomena such as social networks, where the nodes represent people and edges represent friendships. Partitioning nodes into “clusters” such that two nodes within a cluster are more likely to be connected than two nodes from separate clusters can reveal structures such as communities in social networks. Many clustering algorithms have been proposed. However, popular measures of clustering strength are computationally unfeasible to maximize and algorithms can, at best, compute reasonably good clusterings. More importantly, clustering algorithms will return a partition of the nodes even for graphs lacking community structure. We investigate how to determine whether the returned clusters represent a meaningful structure in the graph. We present an algorithm-agnostic method of examining stability (and thus meaningfulness of clusters): after running a clustering algorithm on the initial graph, we add a single edge not present in the original graph and re-run the clustering algorithm. The distance between the clustering on the original and on the modified graphs is recorded, and this experiment is repeated for a large number of edges not present in the original graph. We hypothesize that the distribution of recorded distances carries information about stability and present experimental data collected on a variety of synthetic and real-world graphs.
Funding Provided by: National Science Foundation - DUE-0736738

Anonymity in Online Communities

Eli Omernick (2013); Mentor(s): Sara Sood

Abstract: With the ever-expanding scope of computer-mediated communication, especially in this age of social media and instant communication, there are some interesting and meaningful questions being raised on how we communicate and the subsequent implications; specifically we have investigated the influence of anonymity on the behavior of Internet users. TechCrunch is a technology news site, which posts articles and allows readers to post comments in response. On March 1st, 2011, TechCrunch switched from the Disqus commenting platform to the Facebook commenting platform, marking the end of condoned anonymity in their online community. They did this in the name of “Troll Slaying,” or the attempt of reducing intentionally negative or destructive user contributions. We looked at trends between the two corpora as wholes as well as between the user groups (which we characterize as having varying degrees of anonymity, from totally anonymous, to a pseudonym, to using users’ real names) within the individual corpora. We evaluated comments in terms of several qualitative (e.g. Readability, Relevance, Word Usage) as well as quantitative (e.g. Comments/Article, Comments/User, Comment Length) metrics.
Funding Provided by: Pomona College SURP

Research at Pomona