Why do newcomers abandon open source software projects?
0 likes963 views
This document summarizes a study on why newcomers abandon open source software projects. The study examined mailing lists and issue trackers for the Hadoop project over 60 months to analyze responses to 67 newcomers. It found that the absence of responses to questions correlated with higher abandonment rates. Newcomers who received polite, on-topic answers were more likely to remain involved with the project. The type of author of the response also influenced retention, with existing project members' responses leading to better outcomes. Interviews with 11 newcomers confirmed that a lack of guidance made it difficult for them to continue contributing. The conclusions were that initial interactions strongly affect a newcomer's decision to stay involved and that "social skills" are important for
![CONTEXT
... and need support on their
first steps
2
Newcomers are explorers who must
orient themselves within an
unfamiliar landscape...
[Degenais et al. 2010]
B. Dagenais, H. Ossher, R.K.E Bellamy, M.P. Robillard and J.P.
de Vries, Moving into a new software project landscape, in
ICSE 2010.](https://image.slidesharecdn.com/chase13present-130625131802-phpapp02/85/Why-do-newcomers-abandon-open-source-software-projects-2-320.jpg)
Why do newcomers abandon open source software projects?
- 1. WHY DO NEWCOMERS ABANDON OPEN SOURCE SOFTWARE PROJECTS? Igor Steinmacher Igor Scaliante Wiese Ana Paula Chaves Marco Aur辿lio Gerosa
- 2. CONTEXT ... and need support on their first steps 2 Newcomers are explorers who must orient themselves within an unfamiliar landscape... [Degenais et al. 2010] B. Dagenais, H. Ossher, R.K.E Bellamy, M.P. Robillard and J.P. de Vries, Moving into a new software project landscape, in ICSE 2010.
- 3. GOAL Research Question 4 Does the absence of response, politeness, usefulness or the author of answers influence the retention of newcomers in an open source project? Specific goals: During newcomers first interactions with the project check if they receive answers; observe who are the authors of the answers to their questions; classify the answers received by the newcomers.
- 4. DATA Hadoop Common Project 60 months Mailing lists 9285 threads 50128 messages Issue tracker discussions (Jira) 7721 issues 76065 comments 5
- 5. Issue Tracker Mailing List Identification of Existant Members (36 months) Newcomers Identification (6 months) Newcomers that Returned (6 months) Retained Newcomers (+ 12 months) Newcomers that Abandoned (6 months) Survey Extractors 6 1 2 3 3 4
- 6. # users % of existant % of newcomers Existant members (interval 1) 677 Newcomers (interval 2) 67 9.90% Remaining (interval 3) 20 2.95% 29.85% Retained (interval 4) 12 1.77% 17.91% RESULTS (RETENTION) 7 Mailing list newcomers # users % of existant % of newcomers Existant members (interval 1) 483 Newcomers (interval 2) 127 26.29% Remaining (interval 3) 30 6.21% 23.62% Retained (interval 4) 17 3.52% 13.39% Issue tracker newcomers (Jira)
- 7. SG1: ARE NEWCOMERS ANSWERED? 8 # people Abandoned Appear on intervals 3 and 4 Answered 13 11 (84.6%) 2 (15.4%) Not answered 34 30 (88.2%) 4 (11.8%) # people Abandoned Appear on intervals 3 and 4 Not commented 7 6 (85.7%) 1 (14.3%) Received comments 71 55 (77.5%) 16 (22.5%) Issue tracker newcomers (Jira) Mailing list newcomers
- 8. 20 threads triggered by newcomers that abandoned the project had some answer from other newcomer SG2: WHO ANSWERS NEWCOMERS? 9 Mailing list No message sent by a remaining newcomer was answered only by other newcomers. There are some hints that the authors of answere influence the newcomers After manual analysis: content of the answers seems to be more relevant
- 9. SG2: WHO ANSWERS NEWCOMERS? 10 Jira Very heterogeneous participation Discussions are contextualized and focused Higher amount of messages exchanged per thread (issue)
- 10. SG3: WHAT KIND OF ANSWER DO NEWCOMERS RECEIVE? Answer type Left the project Remained Useful / On Topic 20 7 Not Useful / Another question 5 0 Not Useful / Offtopic 3 0 Indifferent 1 0 Other 4 0 11 Answers are in topic, contextualized and bring useful information Mailing List Issue tracker
- 11. RESULTS QUESTIONNAIRE 1. Do you remember sending an email to hadoop-common-dev mailing list? 2. At that time, were you interested to keep contributing to Hadoop project? 2a. In case you answered YES to question 2, why did you give up? 2b. In case you answered NO to question 2, what was the goal of the messages sent to developers list? 3. Have you contributed to the project after June 2009? 4. Have you contributed to other Open Source project BEFORE 2009? 12 Question 1 Question 2 Question 3 Question 4 Yes 13 11 1 7 No 0 2 12 6 11
- 12. RESULTS QUESTIONNAIRE Type of answer Answer to question 2 Yes (2a) No (2b) i. The user just wanted to clarify some doubt 0 2 ii. Question not answered or answer did not help 2 0 iii. Lack of help to choose a task 3 0 iv. Not accepted by the Project 1 0 v. Changed focus or company 4 0 vi. Resumed the contributions later 1 0 13 2a. In case you answered YES to question 2, why did you give up? 2b. In case you answered NO to question 2, what was the goal of the messages sent to developers list? ... My issue was how to start contributing If I could have got some hand holding that might have helped I got no answer for my question
- 13. CONCLUSIONS absence of answer politeness of answers receptiveness (survey) usefulness of answer type of author Initial interactions influence newcomers decision to recontribute to the project Social skills are required 14 Does the absence of response, politeness, usefulness or the author of answers influence the retention of newcomers in an open source project?
- 14. QUESTIONS? Igor Steinmacher igorfs@utfpr.edu.br Igor Scaliante Wiese igor@utfpr.edu.br Ana Paula Chaves anachaves@utfpr.edu.br Marco Aur辿lio Gerosa gerosa@ime.usp.br