1. Socio-Technical Evolu0on
of the Ruby Ecosystem in
GitHub
Eleni Constan,nou, Tom Mens
SANER – 22nd February 2017, Klagenfurt, Austria

2. Introduc0on
SoAware ecosystem
• CollecEon of soAware projects that are developed and evolve together in the
same environment [1]
Socio-technical perspecEve
Sustainable ecosystem
[1] M. Lungu. Towards reverse engineering soAware ecosystems. Int'l Conf. SoAware Maintenance, pages 428-431, 2008.
[2] D. Dhungana, I. Groher, E. Schludermann, S. Biffl. SoAware ecosystems vs. natural ecosystems: learning from the ingenious mind of nature. Eur. Conf. on SoAware
Architecture: Companion Volume, pages 96-102, 2010.
Increase or maintain its user/
developer community over longer
periods of Eme and can survive
inherent changes (such as new
technologies or new products) that
can change the popula,on [2]
2

3. Social changes
Stayers
Time 2 Time 1 Time 3
Joiners
Leaver 3

4. Technical Changes
in projects
Ac,ve
Time 2 Time 1 Time 3
New
Obsolete 4

5. Objec0ve
InvesEgate the socio-technical evoluEon of soAware ecosystems
Ruby ecosystem
GHTorrent dataset [3]
• 9 year Emespan: October 2007 – September 2016
• Time unit: year quarters
[3] G. Gousios. The GHTorrent dataset and tool suite. Working Conf. Mining SoAware Repositories, pages 233-236, 2013. 5

6. Analyzed Data
Base Forks Ecosystem
Projects >25.5K >69K >94K
Contributors >59.5K >45K >76K
Commits >3.5M >1.1M >5M
Touched Files >780K >235K >815K
LOC >427M >114M >542M
Filters [4]
• InacEve and
isolated projects
• Short-lived
contributors
6
[4] E. Kalliamvakou, G. Gousios, K. Blincoe, L. Singer, D. M. German, and D. Damian, “An in-depth study of the promises and perils of mining GitHub,” Empirical
SoAware Engineering, vol. 21, no. 5, pp. 2035–2071, 2016.

7. How does the ecosystem grow over 0me?
A"er mid-2013:
Ecosystem stabilizes
2008
2009
2010
2011
2012
2013
2014
2015
2016
2000
4000
6000
8000
10000
Projects
Obsolete Projects
New Projects
Active Projects
7

8. How does the ecosystem grow over 0me?
8
Before 2014:
30-40% new base projects
<10% obsolete
From 2014:
More projects are
abandoned than created
2008
2009
2010
2011
2012
2013
2014
2015
2016
0
0.2
0.4
0.6
0.8
1
ProjectRenewal
ProjectAbandonment

9. How does the ecosystem evolve (social view)?
From 2014:
Ecosystem stabilizes
2008
2009
2010
2011
2012
2013
2014
2015
2016
2000
4000
6000
8000
10000
12000
14000 Team
Leavers
Joiners
Stayers
9

10. How does the ecosystem evolve (social view)?
10
2008
2009
2010
2011
2012
2013
2014
2015
2016
0
0.2
0.4
0.6
0.8
1
TeamRenewal
TeamAbandonment
From 2014:
more Leavers, less Joiners

11. Evidence of contributor migra0on to JavaScript
11
Ecosystem
Ac,ve in Ruby
(Contributors)
JavaScript 28,322
Ecosystem
Abandoned Ruby
(Contributors)
Percentage
JavaScript 20,198 71%
Most Ruby Leavers…
• worked in JavaScript projects in
parallel to Ruby projects
• ConEnued to work in JavaScript aAer
abandoning Ruby
2008
2009
2010
2011
2012
2013
2014
2015
2016
0
100
200
300
400
500
Numberofactiveprojects
×103
Ruby
JavaScript

12. Do important contributors leave the
ecosystem?
Diversity index of Leavers (also
used in ecology)
Increased specializaEon
• Large contribuEon to
important projects
2008
2009
2010
2011
2012
2013
2014
2015
0
1
2
3
4
Specialization
12

13. External valida0on
Ruby gem downloads:
Drop from late 2014 onwards
2008
2009
2010
2011
2012
2013
2014
2015
2016
0
2
4
6
8
10
#ofDownloads
×10
8
RubyGems in GitHub
RubyGems
13

14. Summary
Growth of the Ruby ecosystem unEl 2014
Socio-technical decline from 2014 onwards
Preliminary evidence of contributor migraEon to other ecosystems
14

15. Take-away message/Future work
Major social changes can highly impact the ecosystem evoluEon
Measuring these changes can help in idenEfying such issues early
IdenEfy which external factors affect
ecosystem sustainability and health
15

16. Which mechanisms can
preserve ecosystem health
in the case of major socio-
technical changes?
Which factors impact
ecosystem health?
16

17. Thank you!
eleni.constanEnou@umons.ac.be
tom.mens@umons.ac.be