Filetype PDF | Posted on 30 Jan 2023 | 3 years ago
Authentication
digital resources
342x Filetype PDF File size 0.55 MB Source: www.sc.ehu.es
FOCUS: THE SOFTWARE ENGINEERING PROCESS
iterative, rapid, structured, object-
Choice of oriented, and agile methodologies.
Each has its virtues, and each has its
supporters and critics.
Software In 2003, IEEE Software pub-
lished a special issue about the state
of software engineering and soft-
1
Development ware development. Articles in that
issue offered insights into the in-
ternational use of methodologies
and refl ected on the practices, tech-
Methodologies niques, and tools implemented in
software projects. Since then, the use
Do Organizational, Project, of agile and hybrid methodologies
has grown. It would be pertinent to
discover which methodologies are
and Team Characteristics common today and why organiza-
tions choose a specifi c methodology.
Matter? For practitioners, determin-
ing the specifi c methodology for a
given project is critical. Sometimes,
Leo R. Vijayasarathy and Charles W. Butler, the choice of methodology might be
Colorado State University based on marketing and literature
bias that supports new or industry-
supported practices. At other times,
// Survey results indicate that although companies might rely on standards
agile methodologies are more prevalent for consistency and repeatability.
It’s doubtful that choosing a meth-
than 10 years ago, traditional methodologies odology will ever be a simple de-
are still popular. Organizations also terministic exercise. Rather, the
selection will likely consider sev-
use multiple methodologies on projects. eral contextual factors, including
Furthermore, their choice of methodologies organizational, project, and team
characteristics, as well as market
is associated with certain organizational, and operational forces. So, guide-
project, and team characteristics. // lines drawn from empirical asso-
ciations between the methodologies
used and key situational character-
istics would help support informed
decision making.
Toward that end, we performed a
study to empirically assess the extent
to which different software develop-
ment methodologies, including tra-
ditional, iterative, and agile, are in
use. We also sought to determine the
SOFTWARE DEVELOPMENT meth- the process of developing software associations between the methodolo-
odologies provide a framework for systems. Many methodologies ex- gies and organizational, project, and
planning, executing, and managing ist, including waterfall, prototyping, team characteristics.
86 IEEE SOFTWARE | PUBLISHED BY THE IEEE COMPUTER SOCIETY 0740-7459/16/$33.00 © 2016 IEEE
45
41.8
40
35
30
25
20
Percentage of respondents15
11.8
10 9.8 8.5
6.5
5 2.6 3.3 3.3
0.7 0.7 2.0 1.0 0.7
0
r h r
ester Othe
Analyst T
Developer eam leader Agile coac
T Product owner Project sponsor Scrum master
Project manager
Architect or designe Development manager
Quality assurance manager
FIGURE 1. The respondents’ primary project roles. Managers and leaders accounted for 58.2 percent of the roles; the technical
roles (analyst, architect or designer, developer, and tester) accounted for 25.5 percent.
Online Survey Respondent Profi le Industry Profi le
and Data Profi les Figure 1 shows the respondents’ pri- The respondents’ organizations cov-
We collected the study data through mary project roles. Of the 153 respon- ered many industry sectors (see Fig-
an anonymous online survey. Our dents, 41.8 percent identifi ed them- ure 2). The largest percentage of
university’s Research Integrity & selves as a project manager. Other respondents were from the informa-
Compliance Review Offi ce approved common roles were team leader (11.8 tion technology sector (26.8 per-
our study design and questionnaire. percent), analyst (9.8 percent), archi- cent). Other well-represented sectors
We posted the survey on the Project tect or designer (8.5 percent), and were fi nance, banking, or insurance
Management Institute’s (www.pmi tester (6.5 percent). Managers and (13.7 percent); government or pub-
.org) website, in the Academic Re- leaders accounted for 58.2 percent of lic administration (11.1 percent);
search section. We also sent our sur- the roles; the technical roles (analyst, professional, scientifi c, or technical
vey link with a solicitation message architect or designer, developer, and services (8.5 percent); and the medi-
to 2,000 project managers and team tester) accounted for 25.5 percent. cal, dental, or healthcare professions
members. Our questionnaire used Many respondents had signifi cant (6.5 percent).
Qualtrics survey-building software, experience. Eighty-six (56.2 percent) Respondents indicated their or-
and it was hosted on their site. The reported more than fi ve years’ ex- ganizations’ size in terms of the an-
questionnaire asked participants to perience as a manager or lead. One nual revenue and number of employ-
base their responses on a recently hundred and two (66.7 percent) re- ees, using the ranges of values we
completed software development ported more than fi ve years’ experi- provided. These measures’ median
project in which they had played an ence as an analyst, architect or de- ranges were US$1 to $100 million
active role. signer, developer, or tester. and 251 to 500 employees.
SEPTEMBER/OCTOBER 2016 | IEEE SOFTWARE 87
FOCUS: THE SOFTWARE ENGINEERING PROCESS
30
26.8
25
20
15 13.7
11.1
Percentage of respondents10 8.5
5.9 6.5
4.6 5.1 5.1
5
2.0 2.0 2.7
1.3 1.3 0.7 0.7 0.7 1.3
00
0
e n g s g l s y r
vice vice
Education Mining tilities the
Agricultur Construction ublishin
Manufacturin
lecommunication
holesale or retai
e efense or militar
Hotels or food ser
Information technology
Finance, banking, or insurance Medical, dental, or healthcare ansportation or arehousing
r
Art, entertainment, or recreation eal estate or rental or leasing
Government or public administratio
rofessional, scientific, or technical ser
FIGURE 2. Industry distribution. About 40 percent of our respondents were from the information technology and financial, banking,
or insurance sectors.
The industry sectors had a nar- Figure 3b shows the project dura- development project budgets were
row geographic distribution. The tion. Projects generally were short, $400K or less, and 70 percent of
US accounted for 133 respondents with 76.4 percent lasting 12 months software enhancement project bud-
(86.9 percent). Other countries in- or less. Most of those projects took gets were $400K or less.
cluded India, China (including Hong less than six months. Only 16.4 per-
Kong), the UK, Germany, Romania, cent of the projects took 13 to 24 Methodologies
Sri Lanka, France, Singapore, and months, and 7.2 percent of the proj- and Approaches
Saudi Arabia. ects took more than 24 months. In Figure 4 shows the methodologies
total, 92.8 percent of the projects the respondents used, chosen from a
Project Profile took 24 months or less. list of common methodologies rang-
The respondents worked on a vari- Only 3.9 percent of the projects ing from generic frameworks (for
ety of projects (see Figure 3a). The had budgets of more than $1M. In example, Scrum and PRINCE [Proj-
most common project types were new contrast, 76.4 percent had budgets ects in Controlled Environments])
software development (38.6 percent), of $400K or less. Very few budgets to more specialized processes or
software enhancement (24.2 percent), exceeded $400K for project types techniques (for example, feature-
customization of commercial off-the- other than new software develop- driven development and Extreme
shelf software (13.1 percent), and ment and software enhancement. Programming). Surprisingly, the
software integration (9.8 percent). Even so, 77 percent of new software most frequently used methodology
88 IEEE SOFTWARE | WWW.COMPUTER.ORG/SOFTWARE | @IEEESOFTWARE
45 50
40 38.6 45 43.1
35 40
30 35 33.3
25 24.2 30
20 25
Percentage of projects15 13.1 20
9.8 Percentage of projects15
10
6.5
3.9 10 9.2
5 2.0 2.6 7.2
0 5 3.3 2.6
(a) e t n n r (b) 1.3
Othe 0
System migratio
Software enhancement Software integration 6 to 12 mos.
Less than 6 mos. 13 to 18 mos. 19 to 24 mos. 25 to 30 mos. 31 to 36 mos.
New software development More than 36 mos.
Outsourced software developmen
Enterprise-resource-planning implementatio
Customization of commercial-off-the-shelf softwar
FIGURE 3. Project (a) type and (b) duration. Over half the projects developed new software or enhanced existing software; three-
quarters of the projects were completed within a year.
was the waterfall model (32.0 per- remainder of the project. So, we cat- software development, Extreme Pro-
cent). Other popular methodologies egorized the projects by software de- gramming, Crystal, and dynamic
were Agile Unified Process (28.1 velopment approach rather than in- systems development. The iterative
percent), Scrum (20.3 percent), and dividual methodologies. segment includes projects that used
test-driven development (19.6 per- Figure 5 breaks down the proj- Rational Unified Process, Joint Ap-
cent). None of the methodologies ects by approach. The traditional- plication Development, and rapid ap-
was used in more than one-third of approach segment includes projects plication development. Finally, the
the projects. that adopted one or more plan-driven, hybrid segment includes projects
Perhaps our most interesting find- sequential methodologies such as the that blended methodologies from the
ing was the prevalence of a hybrid waterfall model, the structured sys- other segments.
approach to methodologies; projects tems analysis and design method, These results are similar to those
frequently used multiple methodolo- and PRINCE. The agile segment in- of prior research (see the sidebar).
gies. For example, one respondent cludes projects that used Agile Uni-
cited using Joint Application De- fied Process, Scrum, test-driven devel- Methodology Indicators
velopment for requirements gather- opment, feature-driven development, Researchers have argued that there’s
ing and the waterfall model for the adaptive software development, lean no silver-bullet or one-size-fits-all
SEPTEMBER/OCTOBER 2016 | IEEE SOFTWARE 89
no reviews yet
Please Login to review.
