We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 5
Abstract
Parametric tests are the most common statistical tests applicable when the outcome
variable is continuous in nature. T-test and ANOVA are applicable when independent
comparisons are made between two and more than two independent groups,
respectively. Similar comparisons between paired observation or related samples are
done through their alternatives—paired t-test and repeated measures ANOVA. The
application and interpretation of these parametric tests are discussed in this chapter. Post
hoc tests are performed if the results of ANOVA and between-subjects RMANOVA are
significant. We discuss the application and interpretation of post hoc tests. We also
explore the differences among ANOVA, ANCOVA, MANOVA, and MANCOVA.Introduction
A Mobile Application, or App, refers to software run on mobile phones or smart devices.
Millions of Apps are available via App stores like Google Play* and Apple App? Store [1].
App downloads are projected to increase to 258.2 billion in 2022 from 178.1 billion in
2017 [2]. The pervasiveness of App use also means that quality becomes a major concern.
Fierce competition [3] means that a reliable App will be more successful. While Apps
share common technology with other software, especially web applications, they differ
from desktop software in some important ways [4]: interaction with other applications;
sensor handling such as touch screens and cameras; both native and mobile web
applications; a multitude of hardware devices and platforms; heightened security
concerns; usability that is influenced by other Apps and by the common small size of the
smart phone; power consumption; and complexity of testing.
The complexity of testing arises from the fact that, in addition to the same issues as
found in web applications, App testing must deal with issues related to mobility,
transmission through software, and the issues listed above. Testing mobile Apps is clearly
more complex than testing desktop applications [5]. Muccini et al. [6] investigated how
mobile App testing differs from testing traditional applications. Mobile connectivity
needs to be tested for different connectivity scenarios, networks, resource usage and
associated performance degradation possibly resulting in incorrect system functioning.
All of these items need to be evaluated, as does energy consumption. Varying device
screen resolutions, dimensions, etc., affect usability requiring usability testing. The large
combination of platforms, operating systems, diversity of devices, and rapid evolution is
challenging for a tester, as it can lead to test explosion. Performance assessment is
crucial. Many of these testing needs require that a functional test be executed for a
number of specific environmental scenarios, set-ups, and devices.This is one reason why test automation is clearly desirable and has been pursued quite
successfully [7], [8], [9], [10], [11], [12], [13], [14], [15]. In most cases, the tools are not
based on a model-based testing approach and still require the development of a test suite
up front. They capture test inputs and play them back, or simply automate existing tests
for different configurations, devices, and platforms.
Our interest is in the model-based black-box testing of mobile applications. Specifically,
we are interested in extending an existing technique, FSMWeb [16] to test mobile Apps.
FSMWeb [16], [17], [18], [19] is a widely cited approach that tests web applications.
Andrews et al. [16] proposed FSMWeb as a black-box model-based testing approach. The
model consists of a hierarchical collection of FSMs. In addition, Andrews et al. [20] study
the scalability issues of traditional FSMs of web applications compared to FSMWeb.
FSMWeb compresses inputs using a special purpose input constraint language [20]
reducing the model by as much as 90% . The case studies [20] show that FSMWeb is more
efficient than conventional FSM techniques. Ran et al. [21] defined input selection for
FSMWeb. Andrews et al. [17], [22] also propose an approach for selective regression
testing of web application using FSMWeb and develop a cost-benefit trade-off
framework between brute force and selective regression testing.
This paper is organized as follows: Section 2 describes existing work related to black-box
testing approaches in testing mobile apps, and the original FSMWeb approach. Section 3
presents the extensions to the FSMWeb approach for testing mobile application
(FSMApp). Section 4 compares FSMApp with another approaches for MBT for mobile
apps [23] using one small example app. Section 5 describes a number of case studies to
compare FSMapp this other approach and explores applicability, scalability, effectiveness,
and efficiency. Section 6 draws conclusions and suggests further work.Section snippets
Background
The background section first summarizes Black-Box Model-Based Testing (MBT)
techniques to test web applications; then it explores existing work for Black-Box MBT
approaches for testing mobile Apps.
Testing process for mobile apps
Our approach to testing Mobile Applications using Finite State Machines (FSMApp) is an
extension of FSMWeb [16] which is described in this section. Fig. 1 shows the phases of
the FSMApp process. FSMApp proceeds in four phases (FSMWeb has only three): Phase 1
builds a hierarchical model HFSM, Phase 2 generates tests from the HFSM, Phase 3 selects
the inputs and Phase 4 compiles and executes tests through automated mobile testing
tools...
Example used to illustrate approach
We illustrate our approach using the Family Medicines List...Comparing FSMApp and ESG
In Section 2, we identified one approach [23] that also performs Black-Box MBT for
Mobile Apps. In this section, we compare FSMApp with this approach [23]. Later in
Section 5, we perform a number of case study comparisons.
We used the Family Medicine App (Section 3.2) to compare both approaches. Section 4.1
introduces the ESG method [23]. Section 4.2 compares the results for both approaches...
Case study objectives
Now, we would like to compare and validate FSMApp and ESG for a larger number of
mobile applications. The case studies cover mobile apps from different domains and with
different sizes. We propose to investigate the applicability, scalability, efficiency and
effectiveness of FSMApp for testing mobile applications. Furthermore, we want to know
how FSMApp compares to ESG in these evaluation areas...