Software release planning (SRP) is the problem of selecting which features or requirements will be included in the next release or releases. It is a crucial step in software development, which happens to be extremely complex given the need to reconcile multiple decision making criteria, (e.g., business value, effort and cost), while considering several constraints (e.g., feature precedencies, resource availa-bility). For this reason, several SRP models have been proposed in the literature. The objective of this study is to provide an updated review of SRP approaches reported in the literature.

1. A Survey on Software Release
Planning Models
David Ameller, Carles Farré, Xavier Franch,
Guillem Rufian
PROFES 2016 - November 23th

2. 2srp survey @ profes 2016 : slide #
▶ Background
 SUPERSEDE Project
 Software Release Planning (SRP)
▶ Approach
 Goal
 Research method
 Selection of studies
 Threats to validity
▶ Discussion of the results
▶ Conclusions
Agenda

3. 3srp survey @ profes 2016 : slide #
▶ Call: H2020-ICT-2014-1
▶ Funding scheme: Research
and Innovation
▶ Proposal number: 644018
▶ Duration (months): 36
▶ Start date: 1st May 2015
SUPERSEDE Project
Goal: To provide methods and tools to support
the evolution and adaptation of software
services and applications by exploiting end-
user feedback and big data
Software Release
Planning

4. 4srp survey @ profes 2016 : slide #
The problem of finding the best combination of
requirements to implement in a sequence of
releases.
▶ SRP seeks to maximize some variables like:
 business value
 stakeholder satisfaction
▶ while satisfying some constrains like:
 dependencies among features
 budget
 capacity
Software Release Planning (SRP)

5. 5srp survey @ profes 2016 : slide #
[purpose] Find and characterize
[issue] the proposed models in the academic
literature
[object] for software release planning
[viewpoint] from the viewpoint of project managers
and software developers.
Our Initial Goal

6. 6srp survey @ profes 2016 : slide #
▶ “A systematic review on strategic release
planning models”
 Svahnberg M, Gorschek T, Feldt R, Torkar R,
Saleem SB, Shafique MU.
 Information & Software Technology 52(3), 2010
▶ Systematic Literature Review:
 28 relevant studies, 24 being models for SRP
 2003-2008, except one from 1997
 16 models from the EVOLVE family (Günther
Ruhe et al.)
Antecedents: Svahnberg et al., 2010

7. 7srp survey @ profes 2016 : slide #
[purpose] Find and characterize
[issue] the proposed models in the academic
literature from 2009 onwards
[object] for software release planning
[viewpoint] from the viewpoint of project managers
and software developers.
Our New Goal

8. 8srp survey @ profes 2016 : slide #
1. Base our Research Questions on those in
(Svahnberg et al., 2010)
2. Forward snowballing from (Svahnberg et al., 2010)
 models published from 2011 onwards
3. Backward snowballing from the papers found in 2.
 models published from 2009 onwards
4. Expert consultation
 extra models
Research Method

9. 9srp survey @ profes 2016 : slide #
RQ1. What SRP models have been presented
since 2009?
RQ1.1. What are the main motivations for the
models?
RQ1.2. What are the inputs processed by the
models?
RQ1.3. What are the outputs generated by the
models?
RQ1.4. What are the algorithms or techniques
applied by the models?
Research Questions (1/2)

10. 10srp survey @ profes 2016 : slide #
RQ2. To what extent have the SRP models
surveyed in RQ1 been validated?
RQ2.1. Are the models supported by tools?
RQ2.2. How has industry been involved in the
models?
RQ2.3. What are the major threats identified on
the models?
Research Questions (2/2)

11. 11srp survey @ profes 2016 : slide #
Selection of studies
1st Insertion Criteria (1IC): Relevant & Quality venues
2nd Insertion Criteria (2IC): A SRP model is presented
Papers
citing
the SLR
2IC
Papers cited
by the 2256
103
22
9
647
240
125
6
101
1IC
Scopus Google Scholar
Papers > 2008
9 26
17
Experts
Forward Snowballing
Backward
Snowballing

12. 12srp survey @ profes 2016 : slide #
2009 2010 2011 2012 2013 2014 2015 2016
SRP Models found by year
M2 M1
M4
M3
M5
M6
M8
M9
M7M10
M14
M11
M15
M12 M13
M17
M16
EVOLVE family (G. Ruhe)

13. 13srp survey @ profes 2016 : slide #
▶ Construct Validity. Snowballing narrows the search
scope to the referenced papers, therefore some
papers may be left out.
 Mitigated by the fact that started from a SLR
published in a main software engineering journal
and such SLRs are normally cited by many
researchers
▶ Internal Validity. Each paper was analyzed in depth
by one researcher of this study;
 papers were checked by a second researcher
when doubts arose.
Threats to validy (1/2)

14. 14srp survey @ profes 2016 : slide #
▶ External Validity. As in most literature reviews, this
study does not aim to generalize results because
there is no statistical basis to claim that the selected
papers are a representative sample of the
population
▶ Conclusion Validity. As in any other literature
review, we relied on the result of search engines
which may offer different results in the future.
Therefore, a replication of this study could lead to
different selection of primary studies, and thus to
different results.
Threats to validy (2/2)

15. 15srp survey @ profes 2016 : slide #
RQ1.1. What are the main motivations for the
models?
▶ Pursuing scale: In contrast with the results of
(Svahnberg et al., 2010), more models are
aimed to scale in presence of large sets of
requirements
Discussion of the results (1/7)

16. 16srp survey @ profes 2016 : slide #
RQ1.2. What are the inputs processed by the
models?
▶ Most models consider:
 Dependencies among requirements
 Soft factors (e.g. stakeholder preferences or
business value)
▶ However,
 Only 2 models addressed time constraints (*)
 Only 1 model considers the required skills that
requirements need from developers (*)
(*) SUPERSEDE must-haves
Discussion of the results (2/7)

17. 17srp survey @ profes 2016 : slide #
RQ1.3. What are the outputs generated by the
models?
▶ Simple outputs.
 Most models produce a "binary" yes/no output:
which requirements should be implemented for
the next release(s).
 Only 2 models include requirement
implementation scheduling and resource
(developers) allocation (*)
(*) SUPERSEDE must-haves
Discussion of the results (3/7)

18. 18srp survey @ profes 2016 : slide #
RQ1.4. What are the algorithms or techniques
applied by the models?
▶ SRP as a multi-objective problem. Most
modes recognize the existence of different and
often conflicting objectives that need to be
reconciled in the planning of releases.
Discussion of the results (4/7)

19. 19srp survey @ profes 2016 : slide #
RQ2.1. Are the models supported by tools?
▶ Lack of ready-to-market tools.
 1 commercial tool: ReleasePlanner (EVOLVE
family).
 8 (out of 17) papers did not mention any tool
support.
Discussion of the results (5/7)

20. 20srp survey @ profes 2016 : slide #
RQ2.2. How has industry been involved in the
models?
▶ Scarce industrial contributions.
 Very few authors from the surveyed papers came
from industry, in most cases as providers of a
case study
 Only in one single case the industrial authors
were providing the SRP model.
Discussion of the results (6/7)

21. 21srp survey @ profes 2016 : slide #
RQ2.3. What are the major threats identified on
the models?
▶ Non-optimal consideration of threats to
validity. As much as 7 papers (from 17) with no
mention at all at threats to validity.
Discussion of the results (7/7)

22. 22srp survey @ profes 2016 : slide #
▶ EVOLVE family keeps its prevalence, but to a
lesser degree. (57.1% vs 35.3%)
▶ Only 1 non-EVOLVE author in Svahnberg et al.,
appears again in our study
▶ Soft factors where less considered by models in
Svahnberg et al. (57.1%) than in our study
(88.2%)
▶ Industrial validation was more present in
Svahnberg et al. (56% of the models) than in our
study (23.5%)
Field evolution: comparison with Svahnberg et al.

23. 23srp survey @ profes 2016 : slide #
▶ Literature review on SRP models since 2009
▶ Emphasis on the characteristics of these models
and their validation state.
▶ We have used (Svahnberg et al., 2010) as main
reference to our research methodology.
▶ Positive findings:
 Special attention to the scalability of the models.
 Increasing consideration of soft factors like
stakeholder preferences and business value.
Conclusions (1/2)

24. 24srp survey @ profes 2016 : slide #
▶ Negative findings:
 Incomplete input factors
 Simple outputs
 Proof-of-concept tool support, except for the
case of the EVOLVE family
 Poor industrial validation
 Non-optimal consideration of threats to validity
SRP scientific proposals have not yet reached the
maturity required by industrial contexts
Conclusions (2/2)

25. Questions?
A Survey on Software Release Planning Models