This package contains all the data used in this study. The main file is dataset.csv
, which contains the complete dataset with understandability and naturalness metrics. It contains the following columns:
developer
: ID of the developer who evaluated the snippet;topic
: the topic of the snippet;title
: the signature of the snippet;system_name
: the name of the system from which the snippet was selected;topic_experience
: discretized developers' experience on the topic (values:top
,medium
,bottom
);experience_objective
: objective experience (see below);experience_subjective
: subjective experience (see below);binary_naturalness
: binary developer-based naturalness;naturalness
: continuous developer-based naturalness;global_naturalness
: binary naturalness computed with a global model;correctness
: percentage of correct answers to the questions about the snippet;understood
: perceived understanding of the snippet (values:0
- not understood, or1
- understood);seconds_needed
: seconds needed to evaluate the snippet;deceptiveness
: continuous deceptiveness of the snippet for the developer.
The folder developers-experience
contains two files:
subjective-experience.csv
: self-assessed developers' experience on different topics (5-point Likert scale, between 1 and 5, or 0 if they have no experience at all);objective-experience.csv
: developers' experience on different topics measured in terms of percentage of correct answers to topic-related questions (between 0 and 1).
The folder snippets
contains all the Java snippets used in the context of this study. This folder also contains the file questions.txt
, with the complete list of questions used to evaluate developer-based understandability.