27.03.2014 Views

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Validation<br />

Research<br />

SSTT [13],[14],[15],<br />

[16],[17],[18],<br />

[19],[20]<br />

TABLE II<br />

STUDIES BY RESEARCH TOPICS<br />

Evaluation Solution Proposal<br />

Research<br />

[21],[22] [23],[24],[25],[26],[27],[28],<br />

[29],[30],[31],[32],[33],<br />

[34],[35],[36]<br />

SPLTT [37],[38] [39] [40],[41],[42],[43],[44],[45]<br />

identify different facets within the selected studies. More<br />

details about the type of research as defined by [4] can be<br />

seen in: http://wp.me/p157qN-3i<br />

E. Data Extraction and Synthesis<br />

All the 33 studies were fully read and submitted to a<br />

predefined form to accurately record the information obtained<br />

by the researchers from the primary studies. The form for data<br />

extraction provides some standard information, such as:<br />

• Tool’s name;<br />

• Date of data extraction;<br />

• Title, authors, journal, publication details (if available);<br />

• Prototype information (if available);<br />

• Website (if available); and,<br />

• A list of each conclusion and statement found for each<br />

question.<br />

Based on the research results, inclusion and exclusion criteria,<br />

a set of tools were selected. A brief description presented<br />

ordered by the publication year of SSTT tools can be seen<br />

in: http://wp.me/p157qN-2Z and SPLTT tools can be seen in:<br />

http://wp.me/p157qN-32<br />

IV. OUTCOMES<br />

In this section, each research question is answered by<br />

analyzing different point of views, highlighting the evidences<br />

gathered from the data extraction process. All of these results<br />

populated the classification scheme, which evolved while<br />

doing the data extraction.<br />

Initially, we analyzed the studies distribution regarding to<br />

the research topics. The classification scheme, detailed in Table<br />

II, allowed us to infer that many of the studies are Validation<br />

Research (31%), Evaluation Research (9%) and Solution Proposal<br />

(60%). On the other hand, no Philosophical, Opinion,<br />

and Experience Papers were reported. For this reason, Table<br />

II did not show them.<br />

A. Tools Development and Usage<br />

From the selected tools, most of them were developed in the<br />

academic environment (14 for single systems and 7 for SPL),<br />

while 7 were developed exclusively in industry (6 for single<br />

system and 1 for SPL). The remaining 5 tools were developed<br />

in both environments, academic and industrial (4 for single<br />

systems and 1 for SPL), as detailed in Table III.<br />

[13], [17], and [19], present case studies executed in an<br />

academy context. [21], [22], and [39], describe case studies<br />

in industry. Finally, [14], [15], [16], [18], [20], [37], and [38],<br />

report experiments in order to evaluate the tools. The other<br />

studies did not describe empirical evaluation.<br />

TABLE III<br />

WHERE THE TOOLS WERE DEVELOPED AND USED<br />

Academy Industry Both<br />

SSTT [23],[25],[26],[27],[15],[14],[28],<br />

[29],[30],[22],[31],[16],[36],[20]<br />

[24],[21],[32],<br />

[34],[35],[19]<br />

SPLTT [41],[42],[43],[37],[44],[38],[45] [40] [39]<br />

[13],[17],<br />

[33],[18]<br />

B. Software Product Lines Adaptability<br />

According to the results (see Table II), only 33% of the<br />

single system testing tools can be used to test SPL. The<br />

other ones (Solution Proposal) were implemented to specific<br />

programming languages, techniques and approaches which<br />

cannot be suitable to the SPL context.<br />

Websob [29], Korat [30], CodeGenie [31], JWalk [16],<br />

Smart [32], REST [34] and JUnitMX [35] are tools able to<br />

be utilized in the SPL testing process.<br />

Testing tools with specific purpose such as: test management,<br />

bug reports, security test, can be used in Software<br />

Product Lines testing if the methodology of using it is adapted<br />

to suit the SPL necessities. Commonalities and variabilities<br />

should be considered also.<br />

C. Testing Levels<br />

Many of the analyzed tools have similar functionalities.<br />

Moreover, the extracted functionalities have analogous goals,<br />

so it was possible to group them. This classification matched<br />

the description presented in [6], which defines the following<br />

groups:<br />

• Unit Testing - Tools that test the smallest unit of software<br />

implementation. This unit can be basically a class, or even<br />

a module, a function, or a software component [10].<br />

• Integration Testing - Tools that test the integration<br />

between modules or within the reference in domainlevel<br />

when the architecture calls for specific domain<br />

components to be integrated in multiple systems [10].<br />

• System Testing - Tools that ensure that the final product<br />

matches the required features [37].<br />

• Acceptance Testing - Tools that will be used by customers<br />

during the validation of applications [10].<br />

• Regression Testing - Even though regression testing is<br />

not a test level [6], some tools were developed to work<br />

with it. For this reason, we considered regression testing<br />

as part of the classification.<br />

This classification can be applied not only for single system<br />

testing tools but also for SPL testing tools. The main difference<br />

is that SPL divide Testing according to two activities [10]:<br />

core asset (grouping Unit and integration testing) and product<br />

development (grouping system and acceptance testing). Table<br />

IV details the classification of the tools according to the testing<br />

level plus regression testing.<br />

D. Testing Tools Evolution<br />

In order to identify the evolution of the tools, we constructed<br />

a timeline for SSTT showed in Figure 5. Figure 6 shows the<br />

631

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!