How do Programmers Use the Internet? Discovering Domain Knowledge from Browsing and Coding Behaviors

The Internet is an effective tool for learners to gain new knowledge. Often, people use search engines (e.g., Google) rather than accessing websites directly. People have their search techniques to find specific information. In particular, people with domain knowledge tend to search more efficiently...

Full description

Saved in:
Bibliographic Details
Published in2022 IEEE International Conferences on Internet of Things (iThings) and IEEE Green Computing & Communications (GreenCom) and IEEE Cyber, Physical & Social Computing (CPSCom) and IEEE Smart Data (SmartData) and IEEE Congress on Cybermatics (Cybermatics) pp. 605 - 610
Main Authors Watanabe, Ko, Matsuda, Yuki, Nakamura, Yugo, Arakawa, Yutaka, Ishimaru, Shoya
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.08.2022
Subjects
behavioral data
Encoding
internet privacy
Knowledge engineering
Learning management systems
pattern analysis
Programming
Search engines
Search problems
Social computing
web browsing
Online AccessGet full text

Cover

More Information
Summary:The Internet is an effective tool for learners to gain new knowledge. Often, people use search engines (e.g., Google) rather than accessing websites directly. People have their search techniques to find specific information. In particular, people with domain knowledge tend to search more efficiently than novices. By understanding the gap between people with domain knowledge and novices, the novice can understand the path to becoming an expert. Therefore, in this study, we wanted to know what differences exist in search and programming behavior with and without domain knowledge. In this experiment, we asked a group with and without domain knowledge to solve ten programming problems and collected search logs (input knowledge) and compilation logs (output knowledge). Specifically, the first dataset consisted of 13 participants who had taken a university programming class. The second dataset consisted of 20 participants who had not taken a programming class and had no domain knowledge. We examined differences in search and compilation behavior based on participants' domain knowledge from this data. Since we observed a difference between each group when referring to the correlation coefficient, we performed a binary classification of novice and experienced participants using Random Forest, and achieved an average precision of 0.95, indicating that there were different trends in behavior with and without domain knowledge.
DOI:10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics55523.2022.00034