Ehsan Zabardast, Javier Gonzalez-Huerta, and Francis Palma

1. The Impact of Forced Working-From-Home
on Code Technical Debt:
An Industrial Case Study
SEAA 2022: EUROMICRO CONFERENCE ON SOFTWARE ENGINEERING
AND ADVANCED APPLICATIONS 2022
Ehsan Zabardast, Javier Gonzalez-Huerta, and Francis Palma
ehsan.zabardast@bth.se
javier.gonzalez.huerta@bth.se
francis.palma@lnu.se
https://rethought.se 1
https://bit.ly/3labm5m

2. Authors
2
• Ehsan Zabardast (BTH)
• Javier Gonzalez-Huerta (BTH)
• Francis Palma (LNU)

3. Context & Motivation
• Technical and design decisions impact the introduction of TD.
• The decisions are made by developers, architects, testers, and
others.
image: Freepik.com
3

4. Context & Motivation
• The working environment impacts the
decisions taken by people.
• The communication and coordination of the
members might be a factor that influences
incurring and repayment of TD [1].
4
image: Freepik.com
[1] R. Bavani, “Distributed agile, agile testing, and technical debt,” IEEE
software, vol. 29, no. 6, pp. 28–33, 2012.

5. The Pandemic hit in early 2020!
• Organisations faced a new challenge: how
to continue their work 100% on distance?
• Challenges in communication and
coordination
• Relying on tools (IM & Video Conferencing)
• Working 100% on distance is different
from distributed software development.
5
image: Freepik.com

6. Industrial Case Study
• Goal
• RQ: How the forced Working-From-Home mode
impacted the accumulation of technical debt?
• First 9 months of WFH.
• Sample and Population
• Context: Company developing smart
banking and financial solutions
• Case: Using microservice architecture
• Units of analysis: 5 Components
• Quantitative and Qualitative analysis
6
image: Freepik.com

7. Research Methodology
7

8. Research Methodology
8

9. image: Freepik.com
Focus Group Procedure
• Step 1: Participants introduced themselves.
• Step 2: Researchers presented the results.
• Step 3: Participants answered questions.
• Step 4: Participants and researchers had a closing discussion.
9

10. 10
TD per week Accumulated TD Component Size Growth TD Density
Quantitative
Analysis

11. Quantitative
Analysis
11
TD per week Accumulated TD Component Size Growth TD Density

12. Qualitative
Analysis
12
TD per week Accumulated TD Component Size Growth TD Density

13. Results
• More TD when working from
the office!
• When working from the
office, components C1, C2,
C3, and C5 incur more TD,
i.e., on average, TD density is
higher during 2019.
13

14. Qualitative Analsis
• We held a focus group with the development team to ask
questions.
• Regarding the overall results.
• Regarding the points of interest
14
image: Freepik.com

15. 15

16. 16
Results
• 3 Factors Impacting TD
• Ways of working
• Laws of evolution
• Source code

17. 17
Results
• 3 Factors Impacting TD
• Ways of working
• Laws of evolution
• Source code

18. 18
Results

19. Threats to Validity
• There are other factors that can impact TD
that we might not be aware of.
• The results are only valid for the context and
the case.
• TD is calculated by SonarQube.
• The company is using SonarQube as a proxy to
measure TD.
• We tried to address the issues related to
reliability by involving the developers and
blinded coding.
19
image: Freepik.com

20. Discussion
• The components incurred less TD during forced WFH.
• “A gut feeling is that technical debt should not have
increased during WFH - rather decreased. Sharing screens
and discussing feels more natural when working remotely.”
• Screen sharing and impromptu meetings help
decrease TD.
20
image: Freepik.com

21. Discussion
• TD increased faster during the forced WFH in C4!
• The impact of responsibility diffusion on the growth of TD is highlighted as a
major factor by the developers.
• The gatekeeping solution can also, in the long run, have negative
consequences for the system due to review fatigue [2].
21
TD per week Accumulated TD Component Size Growth TD Density
[2] A. Tornhill, Software Design X-Rays: Fix Technical Debt with Behavioral Code Analysis. Pragmatic Bookshelf, 2018.

22. Take Away
• Changes in the ways of working, e.g., mob programming and screen
sharing, do not cause a faster accumulation of TD.
• However, in components with no clear ownership, the change from
office to WFH causes TD to grow faster.
• 3 categories of factors impacting TD:
• Ways-of-working related
• Laws-of-evolution related
• Source-code related
22
image: Freepik.com

23. Future Perspectives
• Studying the impact of Ownership
and Contribution misalignment on
Technical Debt
• Replicating the study to other
components in the same
organisation and other organisations.
23
image: Freepik.com

24. The Impact of Forced Working-From-Home
on Code Technical Debt:
An Industrial Case Study
SEAA 2022: EUROMICRO CONFERENCE ON SOFTWARE ENGINEERING
AND ADVANCED APPLICATIONS 2022
Ehsan Zabardast, Javier Gonzalez-Huerta, and Francis Palma
ehsan.zabardast@bth.se
javier.gonzalez.huerta@bth.se
francis.palma@lnu.se
https://rethought.se 24
https://bit.ly/3labm5m