Dynamic language bridges and applications in LED patent citation networks

Abstract

A simple strategy is presented for dynamically interpreting remote procedure calls in scripting languages, resulting in the ability to transparently use libraries from both languages in a single program. The protocol described handles complex, nested arguments and object-oriented libraries.

To demonstrate the robustness of `Bifrost`, it will be used in the analysis of a dataset including a network of LED patents and their metadata.

Then, we will present our full general protocol (`Bifrost`). Example implementations show two useful bridges: one between Ruby and Python, and another between two different Python runtimes.

We will first perform a proof-of-concept for matrix multiplication in Ruby (`Grisbr`), showing the performance of a manual bridge to Python very similar to our intended protocol.