Evaluating Architectural Openness in Mobile Software
Platforms

Mohsen Anvaari
IT University of Gothenburg
mohsen.anvaari@gmail.com

Slinger Jansen
Utrecht University
slinger@slingerjansen.nl

ABSTRACT
Software platform developers do not have insight into the
consequences of architectural decisions made in regards to
openness, while openness determines whether and how third-party
developers adopt the software platform. For mobile
communication devices, however, the adoption by third-party
developers is essential for the growth of an ecosystem, which has
proven to be a successful strategy for mobile communication
device ecosystems. In this research the openness and architectures
of five mobile platforms are compared, to provide further insight
into successful openness strategies for software ecosystem growth.
Keywords
Mobile Software Platforms, Openness Strategy, Platform
Architecture, Platform Accessibility, Qualitative Interview.
1. INTRODUCTION
Open source or proprietary; which of them is more successful?
Which strategy has received more attention from the developers?
Which does lead to more innovative applications? One can say the
open source strategy is more successful due to its creativities and
inventions [1] and on the other hand, one can believe that the
proprietary strategy is more successful because it will lead to more
qualified products due to strong control. The reality, however, is
not that black and white, especially when it comes to the software
on the smartphones called mobile software platform. Considering
a platform as open or closed is rarely a binary decision and
generally is a question of “how open” [2]. The answer to the
question deals with openness strategy of the platform.
Openness strategy is the degree to which a platform supplier
allows the platform users to interact with the platform, view,
extend or change its components and depends on different
technical and commercial aspects such as platform architecture,
platform accessibility, platform transparency, licensing state,
marketing policy, etc. Mobile software platform means the overall
structure of the software on the mobile devices [3]. The openness
strategy is different among various mobile software platforms of
smartphone ecosystems. A software ecosystem is “a set of actors
functioning as a unit and interacting with a shared market for
software and services, together with the relationships among
them” [4]. When the definition comes to the smartphone area
software platform suppliers, device manufacturers, operators,
application developers, device customers, etc are considered as
the actors and participants of the ecosystem.
In the current smartphone ecosystem Symbian, Windows Mobile,
iPhone, BlackBerry, and Android are the main software platforms
competing for superiority [5]. Some of these mobile platforms are
more successful than others and many differences between these
platforms exist. Some are available for any hardware such as the
Android platform, whereas others are only available on limited
hardware such as the iPhone platform. To implement the openness
strategy that platform suppliers have made, the architecture of
their platform should support proper platform accessibility for
their application developers and device manufacturers.
This research studies the openness strategy of the different mobile
platforms. The scientific contribution of this research lies in the
identification of the openness strategy of the main mobile
platforms based on their architecture. The platform suppliers
define their openness strategy based on their business model by
considering the architectural aspects of the platform. The strategy
they choose affects device users, application developers, device
manufacturers and operators. In this research, the architectural
aspects of the platforms and some licensing aspects related to
platform accessibility are studied. Platform architecture is the
structure of the software platform compromises its components
and the relationship between them [6] and platform accessibility
means the methods and points that developers can use to extend
or modify the platform. Since other aspects of the business model
such as marketing are not in the scope of software engineering
they are not studied in this research. To validate the openness
strategy in different mobile platforms, only application developers
are chosen to be interviewed. This is due to time limitation,
although the device manufacturers are also in the scope of this
research.
The remainder of this paper is structured as follows: In section 2,
the research questions of the study are clarified and a summary of
the research methods is presented. Before describing the activities
of the study, a summary of related works is presented in sections 3
and in section 4 results of the study are described. Section 6
describes the openness strategy in main mobile platforms.
Sections 6 and 7 discuss the research results and argue how the
research methods provide an answer to the research questions.
Part 7 summarizes all sub-questions and answers, and provides
some pointers for future research.
2. RESEARCH APPROACH
The research question and sub-questions are defined as follows:
RQ: How does the software architecture expose the openness
strategy of mobile software platforms?
• SQ1: Can a model be developed that describes the
architectural openness of mobile platforms?
• SQ2: How would the licensing aspects of the platforms
relate to such a model?

• SQ3: How open are the five main mobile platforms?
In the following section the methods chosen for this research are
discussed.
To answer above questions, literature reviews of architectural
descriptions and qualitative interviews with third-party developers
are held. The literature review has the aim of finding the
relationship between openness strategies and the software
architecture of mobile platforms. Furthermore, the literature study
aims to find out how application developers can extend the
platform. The second research method is that of qualitative
interviews. These semi-structured interviews with mobile
application developers are held to verify the platform accessibility
in the five main platforms that enables application developers to
extend the software platform, which previously has been captured
from the literature and platform documents. An architectural
reference model is created to connect between the openness
strategy and the software architecture of a mobile software
platform.
In qualitative researches, the two main types of interviews are
semi-structured interview and unstructured interview.
“Researchers sometimes employ the term qualitative interview to
encapsulate these two types of interviews” [6]. For this research
semi-structured interviews have been chosen. The main objective
of this study is validating the openness strategy of mobile
platforms by gaining experiences of mobile platform developers.
To prepare the questionnaire, some general questions were created
based on the objectives of interviews. Some more specific
questions about the favorite mobile platform(s) of interviewees
were added to the list of questions. Simultaneously, several
application developers in the Netherlands were contacted to see
whether they are interested to be interviewed or not. The plan was
to find one application developer for each platform. After
finalizing the questionnaire and making appointments with
interviewees, the interviews are conducted. The interviews are
recorded by a voice recorder and all interviews have been
transcribed. Finally for analyzing the interview data, the proposed
stage-by-stage method by Burnard is applied. The aim of this
method is “to produce a detailed and systematic recording of the
themes and issues addressed in the interviews and to link the
themes and interviews together under a reasonably exhaustive
category system” [8].
3. RELATED WORK
The most recently published article in the area of smartphone
ecosystem states that few attempts have been made to demonstrate
the comparison of smartphone OSs [9]. It is also true particularly
about comparing the openness strategies in different mobile
software platforms. Lin and Ye, themselves, have compared the
ecosystem of iPhobe, Symbian, Windows Mobile and Blackberry
by discussing about the value chain (they call it “food web”) of
the ecosystems, but they have not discussed the openness strategy
of the platforms based on the architectural aspects. Further,
Android is not in their comparison. Cho and Joen have discussed
and compared Symbian, Linux and Rex [10], which, except
Symbian, are not the current major platforms. Besides that,
although they have compared the architecture of the platforms and
also the openness strategy of the platforms, but not a connection
between two subjects has been settled. Yamakami has covered
Android, Symbian, iPhone and Windows Mobile in his
competitive analysis [10]. But the factors he has considered in his
analysis are governance, ease of maintenance, interoperability,
ecosystem, cost reduction and reliability, which are not related to
architectural aspects. Constantinou has discussed open source in
mobile platforms and shows the current state of openness in
software stack of mobile platforms [12]. It is very general and not
applied to any specific platform.
Beyond the software platforms on mobile devices, in the software
area in general, there is again a lack of works that discuss how
software architecture is treated in open source projects [13].
Nakagawa et al. have shown that software architecture has an
important role in open source projects via a case study, but they
have not discussed how the openness strategy affects the
architecture. Prehn has concluded in his article that one
characteristic that is shared by the largest and most successful
open source projects is a software architecture and it has to be
very modular [14]. Arief et al. in a similar conclusion shows that
for open source projects the architecture must be modularized
[15]. Both are not applicable results for the case of comparing the
openness strategy in different mobile software platforms.
4. RESULTS
Openness is the “quality or condition of being open” [16]. The
software openness, in the scope of this research, means the degree
to which a software platform approaches to open characteristics
which depend on accessibility and licensing as key aspects. There
might be other aspects which determine openness in a software
platform but only these two are researched here as explained in
the introduction part. Alspaugh et al. [17] in their research
mention several software elements included in common software
architectures that affect the openness of architecture. The elements
are software source code components, executable components,
application program interfaces (APIs) and configured system or
sub-system architectures [17]. The ways that architectural
elements of a software system affect the openness of the system
are being expressed as software extension mechanisms and
approaches. Klatt defines software extension mechanism based on
extension points. An extension point is the definition of the
provided interface for extensions. An extension itself is an
implementation according to an extension point. An extension
mechanism includes everything about an explicit extensible part
of a software” [18]. Jansen et al. mention several mechanisms in
their article that extend software functionality. The mechanisms
are component calls, service calls, source code inclusion, and
shared data objects [19]. To analysis more specifically the
extension in mobile software platforms, a discussion of extension
mechanisms on operating systems is needed since software
platforms on mobile devices are expressed as operating system of
the devices [20]. Alexandrov et al. present different approaches
for extending the operating system. The ways they have
mentioned are extension or modification of different levels of an
operating system from the lower levels to the higher levels. The
approaches are changing operating system (kernel) itself,
modifying device drivers, installing a network server, adding user
level Plug-Ins, making changes to user level libraries, applications
specific modifications and intercept system calls [21]. Their point
of view about extension approaches is employed to develop the
architectural openness model for this research and is discussed in
the following chapters

4.1 Mobile Software Platforms and Their
Architecture
In the last decade, mobile phones have become programmable
handheld computers which have internet connectivity,
computing power and open application programming
interfaces (APIs) providing prospective platforms for an
infinite set of new mobile services and applications [20]. The
new mobile phones which are usually called smartphones have
both hardware and software parts the same as all computing
systems. The software part is called mobile software platform.
Some authors like Verkasalo use software platform as a synonym
to operating system of mobile devices [20]. Some others like Cho
and Jeon believe that software platform of a system means the
overall structure of the software on the system and operating
system is a part of it [10]. Cho and Jeon consider a layered
architecture for the software platform of typical mobile devices
consists of operating system layer, middleware layer and
applications layer [10]. Layered architecture is an architectural
pattern helps to structure systems that can be decomposed into
groups of subtasks in which each group of subtasks is at a
particular level of abstraction [22]. The software platform of
mobile devices is such a system, therefore the layered architecture
is applicable to the architecture of mobile software platforms. The
suggested model for architecture of mobile software platforms for
this research is the same as proposed architecture by Cho and Jeon
with some modifications. In their model, they have not separated
the default applications which are set by device manufacturer
from the applications developed by third-party community and
installed by users. Since third-party applications have a main role
in defining the openness of a platform, the proposed model for
this research has two different sub-layers in the applications layer:
native applications and extended applications. Native applications
are those developed by device manufacturers and in some
platforms are not modifiable. Extended applications are those
developed by application developers and installed by device users,
so these applications extend the applications layer of the platform.
Middleware layer consists of main libraries and services of the
platform like data storage, virtual machine, multimedia libraries,
etc. When application developers create extended application,
they usually call this layer in their applications instead of calling
the core libraries of the platform. The kernel layer, which in Cho
and Jeon’s model is called operating system layer, is the core the
platform. It consists of the lower level components of the platform
such as device drivers, power management framework, security
framework, etc. Figure 1 shows the proposed architecture for
mobile software platforms.
The architecture of main mobile platforms include Android,
iPhone, Symbian, Blackberry and Windows Mobile is looked by
the lens of this architecture and the results are presented in the
following parts. In the next section, the architectural openness
model which is built based on the proposed architecture is
discussed.
4.2 Architectural Openness Model and
Factors
As presented in the previous part, a mobile software platform like
other software systems has an architecture which is the structure
of the platform. A proposed mobile software architecture that is a
general model and can be applied to different mobile platforms
was shown. To discuss the openness strategy of mobile platforms
based on their architecture, the proposed model is not sufficient
and it should be expanded to an “architectural openness model” to
demonstrate platform extension mechanisms and platform
accessibility since these notions have a connection with the
openness concept as discussed before.
The architectural openness model to accommodate the platform
accessibility and platform extension methods, and in a higher
view the platform openness, should illustrate how much and under
which conditions the platform extenders (application developers,
device makers, customers…) can access to different layers and
components of the platform and extend its functionality. Two
online resources of Google Android have used and defined
integrate, extend and modify concepts to clarify the openness
notion in the architecture of Android platforms [23][24]. Sim et
al. mention similar meanings when consider integration and
customization as issues in mobile operating systems [25]. To
expand the “mobile software architecture” presented in the
previous part to “architectural openness model”, the same terms
and definitions are used here:
Integrate a layer: To use the existing components of a layer in a
mobile application via API, Service Call, source code inclusion,
shared data object and other software extensions mechanisms.
Extend a layer: To enhance the functionality of the components
of a layer. The application uses the built-in Google map
application and adds its own functionality on top of Maps is an
example.
Modify a layer: To replace or change the components of a layer.
Writing your own device driver is an example.
The architectural openness model to support the openness strategy
in mobile software platforms is illustrated in the Figure 2.
Figure 1. Architectural Openness Model for Mobile Software
Platforms
Although the model shows the platform access and extension
methods in different levels, but to demonstrate the openness
degree of a platform, licensing aspects of mobile platforms should
also be considered. Different mobile platforms have various
licensing considerations. Besides the possibility to integrate,
extend or modify the components of a platform, the permission of
these activities depends on the platform supplier’s licensing
policy. In some platforms you don’t need any permission to
extend the platform whereas in others you need to submit the
change to the platform supplier first and if they accept, the
changes will be confirmed.
Based on the proposed architectural openness model and possible
states of licensing, the Table 1 shows the architectural openness
factors for mobile software platforms. Applying this table to a

particular mobile platform, the result for instance for middleware
layer of the platform can be like this: Integrate the components of
the layer is allowed and doesn’t need any permission. Extend the
components of the layer is allowed but needs permission from the
platform supplier. And modify the layer is allowed only for some
components and needs permission.
These factors and their statuses provide a continuum to determine
how much a platform is open based on the architecture of the
platform. A platform is considered entirely open if in all the
architectural layers of the platform integrate, extend and modify
the components are possible without any permission from the
platform supplier. On the other end, a platform is totally closed if
within all layers of the platform integrate, extend or modify the
components are not possible. Other platforms are situated between
these two boundaries. In the following part, the situation of five
main mobile platforms being Android, iPhone, Symbian,
Blackberry and Windows Mobile in the openness continuum
based on the architectural openness model and the openness
factors is discussed.

Table 1. Architectural openness factors for mobile software
platforms
Layer Factor Possibility
If possible

Licensing
status
Extended
applications
Integrate extended
applications
Possible/
Possible for
some
components/
Not possible
Permission is not
needed/ In some
situation
permission is
needed/
Permission is
always needed
Extend extended
applications
Modify extended
applications
Native
applications
Integrate native applications
Extend native applications
Modify extended
applications
Middleware
Integrate middleware
Extend middleware
Modify middleware
Kernel
Integrate kernel
Extend kernel
Modify kernel

4.3 Openness in a Mobile Platform from the
Application Developers’ Point of View
The main objective of qualitative interviews in this research is
verifying the openness strategy of main mobile platforms based on
the interviewees’ experiences. The main results of the interviews
are presented in the next section where openness strategy of main
mobile platforms is discussed. In the following sub-sections the
openness in a mobile platform, as a general concept, from the
interviewees’ point of view is discussed which is the implicit
result of the interviews. First, the opinion of interviewees about
the relationship between openness in a mobile platform and
architectural and licensing aspects of the platform is presented.
Then the importance of openness in a mobile platform for
application developers is discussed.
4.3.1 Relationship of Openness in a Mobile Platform
with Architectural and Licensing Aspects of the
Platform
All of the interviewees believe that architectural and licensing
aspects of a mobile platform affects its openness degree. One of
them counts the programming language, the way libraries of a
platform work, if the frameworks are object oriented or not and
the SDKs provided to extend the platform as the architectural
aspects of the platform and believes that “a lot of openness
depends on the architecture”. Another interviewee thinks that if
the architecture of a platform is messy it will be very hard to make
it open and on the other hand, if the architecture is modular and
has layers and subsystems then it will be easy to open it up. In
their point of view, the licensing aspects also affect the openness
of a platform since to access the different layers of the architecture
of a layer or to modify the components of a platform you need to
have a license.
But in their opinion, the architectural and licensing aspects are not
the only parameters that affect the openness degree of a platform.
“You can do a lot with architecture to open a platform but it is not
the only aspect”, one of the developers believes. He explains that
the available samples and examples for developers, the provided
community and the platform documentation are some other
aspects that affect the openness of a platform.
4.3.2 Importance of Openness in a Mobile Platform
for Developers
Overall, for most of interviewees the openness degree of a mobile
platform is not such important and they care less about the
openness of a platform. Since they are commercial developers, for
choosing a platform the financial aspects of the platform is more
important and determinant for them. One of them believes that
managers should more care about openness and they should be
aware what is allowed to do and what is not allowed. Another
explains that if a platform is open enough to make a lot of money
for him, then the platform is interesting for him.
Nevertheless, openness in higher levels of a mobile platform is a
considerable aspect for interviewees. For example, although
Apple has supported some APIs in the kernel layer of the iPhone
and RIM has not supported any API in the kernel layer of the
Blackberry, one of the interviewees considers the Blackberry
more open than the iPhone, because native applications in the
Blackberry are allowed to integrate but in the iPhone are not. All
of the interviewees are more confident about the openness degree
of higher layers in their favorite platform than the lower layers.
They are not sure about kernel layer and most of them about
middleware layer because they have not tried to integrate, extend
or modify components of these layers. One of them explains that
“the middleware you want to integrate is huge, so … you as a
business developer don’t care about extending or modifying it”.
Most of them believe that even if a platform opens up its
middleware or kernel layer, the people who benefit from the
openness of the platform and are interested to access or modify
the components of the lower layers are device manufacturers not
application developers. All of the interviewees are almost satisfied
by current openness degree of their favorite platforms and just one
of them would like to see a small part of his favorite platform to
see opened up which is Wi-Fi library in Blackberry.

5. OPENNESS IN MAIN MOBILE
PLATFORMS
In this section, based on the architectural openness factors,
documentation of the mobile platforms and results of qualitative
interviews, the openness strategy in five main mobile platforms
being Android, iPhone, Symbian, Windows Mobile and
Blackberry is discussed
5.1 Android
In November 2007, Google formed a group of mobile and
technology companies called The Open Handset Alliance. The
aim of this conglomerate is improving mobile phones to change
the mobile experience for customers by increasing the openness in
the mobile ecosystem. Their first joint project is Android which
has been released officially on October 2008 [26]. Android is a
Linux-based OS that's geared to run on lightweight devices like
mobile phones [27]. Software architecture of Android platform is
a layered architecture. It is possible for a developer to code in any
of the layers, one of the Android developers explains. He clarifies
his opinion about openness in Android more by separating the
situation in literature and in reality based on his experience:
According to the official documents “Android is open source. If a
developer wanted to make changes to something in the Kernel,
they could submit it to Google who manages the project, and
Google would approve the change, and allow it into the
framework. In practice, I think it is difficult for anyone outside of
Google to get changes into the core of the framework”. All of the
interviewees believe that even if in practice Android is totally
open and everyone can modify the source, this openness makes
sense for device manufacturers not commercial developers. In
Table 2 the openness degree of Android in different architectural
layers is shown.
5.2 iPhone
In January of 2007, Apple announced iPhone at the MacWorld
expo in San Francisco [28]. The iPhone OS platform was built on
Mac OS X. Despite its similarity to Mac OS X, there are some
technologies available only on iPhone OS such as Multi-Touch
interface [29]. iPhone software platform is built on a layered
architecture. Apple has published some public APIs for the
iPhone that are the access points of the platform and developers
can integrate them in their applications, as the iPhone application
developer explains. There are also some private APIs which, if
developers use them in their applications, the applications will be
rejected when developers want to publish them on AppStore.
Table 2 demonstrates the openness degree of different
architectural layers in iPhone.
5.3 Symbian
Symbian OS is still at the highest position in the worldwide
smartphone market [5]. Its creator is Symbian Limited founded in
1998 by Ericsson, Nokia, Motorola and Psion. But in 2008, Nokia
acquired all the remaining shares of Symbian Limited [9]. After
releasing the Android, Nokia wanted to compete with it in the
openness degree [28], therefore has formed Symbian Foundation
in 2008. Symbian is a layered software platform. After launching
Symbian Foundation and releasing the source code of the
platform, the openness degree of the platform has been risen up.
But the extension of the platform is still under control of Symbian
Foundation and the same as Android device manufacturers are the
main beneficiaries of openness of the platform. In Table 2 the
openness degree of different layers of the platforms is presented.
5.4 Blackberry
“In 1999, Research In Motion (RIM) introduced the Blackberry
which started as a simple two-way pager, but quickly became one
of the most widespread of mobile computing devices … The
device became so widely adopted, that PC magazine ranked it the
14th most important gadget invented in the past 50 years” [28].
There is no significant information about the structure of the
platform in either in literature or in official documents of the
platform. So the architecture of the platform is get from the
interview with a Blackberry application developer. The same as
other mobile platforms, Blackberry software platform has a
layered architecture. The same as iPhone, Blackberry is strongly
controlled by RIM and developers are not allowed to extend or
modify most components of the platform. The native applications
of the platform are more open than iPhone to be integrated into
the developers’ applications. Table 2 shows the openness situation
in different layers of the platform.
5.5 Windows Mobile
Around the same time the Blackberry was gaining traction,
Microsoft released its first OS targeted at the mobile device
market [28]. Microsoft licenses Windows Mobile to any mobile
phone maker who is interested in launching Windows Mobile on
its device [9], therefore it is not a proprietary OS. Operating
system of Windows Mobile is built based on Windows CE which
is an operating system developed by Microsoft handheld
computers and embedded systems. So the architecture of
Windows Mobile is the same as Windows CE which is a layered
architecture. Developers have permission to extend or modify
some components of the lower layers of the platform such as
device drivers. But native applications are not allowed to be
integrated, extended or modified. The openness degree of
different layers of the platform is shown in Table 2.
6. ANALYSIS
The explicit results of the research process show the different
openness strategies in the main platforms. In theory, the Android
and Symbian platforms are almost completely open since Google
and Symbian Foundation have released the whole source code of
the platforms and device manufacturers, developers and users can
download the source code and do everything they want with the
components of each layer of the platforms. In this case even
Symbian is more open since there is no limitation set by Symbian
Foundation about integrating, extending or modifying of any
components of the platform, where in kernel layer of the Android
there is some components like power management which users are
only allowed to integrate it and not extend or modify. But the
situation in practice is different and the experiences of developers
show that these two platforms are not as open as they seem since
there are some controls governed by Google and Symbian
Foundation which do not allow commercial developers to make
every desired change to the platform and the target people of
releasing the source code are device manufacturers like HTC. In
this case Android is more open since there is less restriction set by
Google for submitting an extension or modification of the
platform. However, even in practice Android and Symbian are the
most open platforms among current main mobile platforms,
although they are not completely open. After these two platforms,

Windows Mobile and Blackberry are situated near together. There
are not enough materials on the documentation of these platforms
to discuss about accessibility to different layers of the platforms,
but according to the experiences of the developers, Windows
Mobile is more open than Blackberry since kernel layer in
Blackberry is almost close and users cannot access to it directly
but in Windows Mobile users even can write their own drivers for
the device. But on the other hand native applications users can do
more with Blackberry than Windows Mobile. In Windows
Mobile, even integrate the native applications is not possible, but
in Blackberry some native applications that have APIs can be used
by users. The licensing situation for both platforms is almost the
same and in both cases there is some situation that people can use
unsigned applications. The big difference here is that Windows
Mobile can be installed on different devices, but Blackberry is a
proprietary software platform that is installed only on Blackberry
devices. So totally the conclusion is that Windows Mobile is more
open than Blackberry. And finally in the spectrum of openness,
after these two platforms the iPhone is situated which the least
open platform among the main mobile platforms. Although
developers can integrate the kernel layer of the platform in their
applications and in this sense it is more open than Blackberry, but
on the other hand developers can integrate several native
applications in the Blackberry which is almost not possible for
any native applications of the iPhone. So in the case of
accessibility of architectural layers, Blackberry and iPhone are
situated nearly in the spectrum but the main thing that distinguish
the iPhone from Blackberry and brings it to the end of the
spectrum is licensing situation which is very controlled for the
iPhone and restriction set by Apple is much more than other
platforms since every application before submitting to the
application store should be signed by Apple and the process of
quality testing of the application is strongly controlled. Table 2 in
Table 2 summarizes the comparison of openness strategy in the
main mobile platforms based on the architectural aspects and
licensing situation of the platforms.
This study has also some implicit results which are mainly
achieved from the interviews with developers. The interviews
show that most of developers, which are typically commercial
developers, do not care about architectural openness of a
platform. For developers the tools and languages are supported to
develop applications, guides and documentation for developing
and financial aspects of the platforms are more considerable.
When they want to consider the architectural openness of a
platform they more care about higher layers of the platform. The
reason is that first of all they think even if the platform is mostly
open, the lower layers provided by the platform supplier work
well and they do not need to spend time to extend or modify it.
The second reason is that if they modify lower layers such as
components of the middleware or kernel, the application users
need also to change the middleware or kernel of their devices
which is not a practical work. So in their point of view increasing
the openness of a platform and releasing the source code of lower
layers make sense for device manufacturers who want to
customize the platform for their own devices. This result brings
some limitation for this study which is argued in the discussion
section.

Table 2. Comparison of Openness Strategy in the Main Mobile
Software Platforms

A
nd
roid

Sy
m
bian

W
ind
o
w
s

M
obile

Blackb
erry

iPh
o
n
e

Factor P L P L P L P L P L
Integrate
extended
applications
Pc Pn Pc Ps Pc Ps Pc Ps Pc Pa
Extend extended
applications Pc Pn Pc Ps Pc Ps Pc Ps Pc Pa
Modify extended
applications Pc Pn Pc Ps Pc Ps Pc Ps Pc Pa
Integrate native
applications Po Pn Pc Ps Np Pc Ps Pc Pa
Extend native
applications Po Pn Po Ps Np Np Np
Modify native
applications Po Ps Po Ps Np Np Np
Integrate
middleware Po Pn Po Ps Po Ps Po Ps Po Pa
Extend
middleware Po Pn Po Ps Pc Ps Np Po Pa
Modify
middleware Po Ps Po Ps Np Np Np
Integrate kernel Po Pn Po Ps Po Ps Np Pa
Extend kernel Pc Pn Po Ps Pc Ps Np Pc Pa
Modify kernel Pc Ps Po Ps Pc Ps Np Np
P = Possibility, L = Licensing Status, Po = Possible, Pc = Possible for
some components, Np = Not possible, Pn = Permission is not needed,
Ps = Sometimes permission is needed, Pa = Permission is always needed

7. DISCUSSION
The first limitation of this research is realized in the literature
review stage. As discussed in related work part, although there is
some literature compares most of studied mobile platforms of this
research together, even regarding the openness of the platforms,
but none of them discuss the openness strategy of the platforms
based on technical aspects such as architecture of the platforms or
software access points of the platform. Even literature about
general software platforms do not discuss the openness of
platforms based on the architectural aspects. The architectural
openness model developed in this study is a tool to aim
identification of the openness strategy in mobile platforms based
on their architectural aspects. Despite this model is based on
typical layered architecture of mobile platforms and is applicable
to all main mobile platforms, but the efficiency of the model to
identify the openness strategy of the mobile platforms would be
confirmed by an empirical study which is not done here due to
restricted time of the research. Another boundary in literature

review part is incompleteness of documents of the mobile
platforms to finding out the openness strategy and access points of
the platforms. It is not easy to contact technical engineers and
architects of the companies like Google, Microsoft, RIM, etc. but
if they accepted to be interviewed, the results of the study would
be more reliable.
The second limitation of this study as shown in results of the
interviews is that architectural openness of the studied platform is
not important for interviewees which are commercial developers.
Since they do not care about the openness especially in lower
layers of the platforms, they have not tried to extend or modify the
components of the lower layers. As a result they are not certain
about accessibility of all layers of their favorite platform and it
could affect the reliability of the results. As interviews show,
device manufactures would benefit more from openness in mobile
platforms since they can customize even lower layers of open
platforms like the Android and Symbian and to some extend
Windows Mobile for using the modified platform in their devices.
So the way to improve the validity of research results would be
conducting interviews with some technical engineers in device
manufacturers if it is not very difficult to contact them.
8. CONCLUSIONS AND FUTURE WORK
The overall aim of this research was identification of openness
strategy in mobile platforms based on the software architecture of
the platforms. This section revisits the specific objectives of this
research, summarizes the findings of the research, offers
conclusions based on the findings, and finally suggest some
recommendations for future works based on the limitations of the
research work.
Research Objective 1: Building a model to describe the
architectural openness of mobile platforms - Architectural
openness model is built based on a typical layered architecture
which is applicable to all main mobile software platforms and
compromises applications layer, middleware layer and kernel
layer. Applications layer itself includes two parts, extended
applications and native applications. Besides the layered
architecture, the model shows three way of accessibility to the
platform for each layer. These ways are integrating, extending or
modifying a layer.
Research Objective 2: Defining architectural openness factors
by considering the licensing aspects of mobile platforms in the
model - Although the model shows the platform access and
extension methods in different levels, but to demonstrate the
openness strategy of a platform, licensing aspects of the platform
should also be considered. By considering licensing situation of
mobile platforms, some factors are defined to identify openness
strategy of platforms and presented in table 2. For example for
kernel layer of a platform, the result of applying the openness
factors can be: Modifying the kernel layer is possible for some
components of the layer but it always needs permission from the
platform supplier.
Research Objective 3: Looking at main mobile platforms by
the lens of developed model and factors to determine how open
the platforms are - The openness strategy of main mobile
platforms include Android, Symbian, iPhone, Windows Mobile
and Blackberry is discussed by applying the architectural
openness model and factors in the architecture and licensing
aspect of the platforms which are gained from the literature. The
results are demonstrated by a figure for each platform. Due to
insufficiency of literature and documents of the platforms, some
factors needs to be confirmed by interviews.
Research Objective 4: Conducting some qualitative interviews
with application developers of the platforms to confirm the
results of previous step - For each platform except the Android,
one qualitative interview with an application developer of the
platform is conducted. The developers are asked to explain about
the openness of their favorite platform based on the accessibility
of each architectural layer of the platform and also licensing
situation of the platform. The results show that for some layers
especially lower ones, commercial developers are not sure about
openness degree of the platform because they don’t care about the
openness in the lower layers and have not tried to extend or
modify them. But about the higher layers and licensing situation
of the platform the results of the interviews are valuable. In some
cases the results confirmed the previous results gained from
documents of the platforms, and in some cases the experience of
developers are different from results of literature. So they believe
that even for platforms like Android and Symbian, the openness
degree in the reality is fewer than what is claimed in the theory.
Finally a comparison of openness strategy in five main mobile
platforms is presented in this research which is based on the
results of looking at the platforms by lens of architectural
openness model and factors and evaluation of results by
conducting interviews.
As argued in the discussion section insufficiency of literature and
documents, lack of time to do more interviews with other
application developers and interview with some engineers from
devices manufacturers, restrictions on access to technical
engineers in platform suppliers such as Google and Microsoft are
the main limitations of this research and affect the reliability of
the results. The architectural openness model and factors are
valuable results of this research. It is recommended, for increasing
the reliability of the model and factors, that some expert people
from software architecture and software openness areas also be
interviewed. For improving the validation of openness strategy of
mobile platforms, interview with technical engineers of mobile
device manufactures is also recommended.
9. ACKNOWLEDGMENTS
Our thanks to mobile application developers for their contribution
in the qualitative interviews. Their efforts have influenced
positively the quality of the results of this research and it was
impossible without their support.
10. REFERENCES
[1] Paulson, J.W., Succi, G., and Eberlein, A. An Empirical
Study of Open-Source and Closed-Source Software
Products, IEEE TRANSACTIONS ON SOFTWARE
ENGINEERING, 30, 4 (2004), 246-256.
[2] Maxwell, E. Open Standards, Open Source and Open
Innovation: Harnessing the Benefits of Openness,
Innovations: Technology, Governance, Globalization 1, 3
(2006), 119-176.
[3] Cho, Y.C. and Jeon, J.W. Current Software Platforms on
Mobile Phone, International Conference on Control,
Automation and Systems. Seoul (2007), 1862-1867.

[4] Jansen, S., Finkelstein, A., and Brinkkemper, S. Business
network management as a survival strategy: A tale of two
software ecosystems. In Proceedings of the First Workshop
on Software Ecosystems. CEUR–WS, vol. 505, (2009)
[5] Canalys research release 2010, Available at:
http://www.canalys.com/pr/2010/r2010021.html, Last
retrieved: 2010-04-21.
[6] Bass, L., Clements, P. and Kazman, R. Software Architecture
in Practice, Second Edition. Addison-Wesley, Boston,
(2003), p 21.
[7] Bryman, A. and Bell, E. Business Research Methods,
Oxford: Oxford University Press, (2007), p 104.
[8] Burnard, P. A Method of Analysing Interview Transcripts in
Qualitative Research, Nurse Education Today 11 (1991),
461-466
[9] Lin, F. and Ye, W. Operating System Battle in the Ecosystem
of Smartphone Industry, In Proc. of 2009 International
Symposium on Information Engineering and Electronic
Commerce, (2009), 617-621.
[10] Cho, Y.C. and Jeon, J.W. Current Software Platforms on
Mobile Phone, International Conference on Control,
Automation and Systems. Seoul, (2007), 1862-1867.
[11] Yamakami, T. Foundation-based Mobile Platform Software
Engineering: Implications to Convergence to Open Source
Software, ACM International Conference Proceeding Series;
Vol. 403, In Proc. of the 2nd International Conference on
Interaction Sciences: Information Technology, Culture and
Human, (2009), 206-211.
[12] Constantinou, A. Mapping open source into mobile: who,
where and how, VisionMobile Ltd., available via:
http://www.visionmobile.com/blog/2008/12/mapping-open-
source-into-mobile-who-where-and-how/, Last retrieved:
2010-02-19
[13] Nakagawa, E. Y., Souza, E. P. M., Murata, K. B., Andery, G.
F., Morelli, L. B., Maldonado, J. C. Software Architecture
Relevance in Open Source Software Evolution: A Case
Study, In Proc. of Annual IEEE International Computer
Software and Applications Conference, (2008), 1234-1239
[14] Prehn, S. Open Source Software Development Process, Term
Paper in AG Software Engineering Seminar SS07 (2007).
[15] Arief, B. Gacek, C. and Lawrie, T. Software Architectures
and Open Source Software – Where can Research Leverage
the Most?, In 1st Workshop on Open Source Software
Engineering:Making Sense of the Bazaar (part of the 23rd
ICSE, 2001) 3–5.
[16] Oxford English dictionary – openness definition, Available
at:
http://dictionary.oed.com/cgi/entry/00332458?single=1&que
ry_type=word&queryword=openness&first=1&max_to_sho
w=10, Last retrieved: 2010-04-07.
[17] Alspaugh, T.A., Asunction, H.U. and Scacchi, W. Analyzing
Software Licenses in Open Architecture Software Systems, In
FLOSS ’09: Proceedings of the 2009 ICSE Workshop on
Emerging Trends in Free/Libre/Open Source Software
Research and Development, (2009), 54–57.
[18] Klatt, B. Software Extension Mechanisms, Fakultt fr
Informatik,. Karlsruhe, Germany, Interner Bericht, (2008),
Available: http://www.bar54.de/benjamin-klatt-software-
extension-mechanism.pdf, Last retrieved: 2010-04-23.
[19] Jansen, S., Brinkkemper, S., Hunink, I., Demir, C. Pragmatic
and Opportunistic Reuse in Innovative Start-Up Companies,
IEEE Software, 25, 6 (2008), 42-49.
[20] Verkasalo, H. Open Mobile Platforms, Modeling the Long-
Tail of Application Usage, In Proc. of Fourth International
Conference on Internet and Web Applications and Services,
(2009), 112-118.
[21] Alexandrov, A.D., Ibel, M., Schauser, K.E. and Scheiman,
C.J. Extending the operating system at the user level: the ufo
global file system, In 1997 Annual Technical Conference On
Unix and Advanced Computing Systems (USENIX’ 97).
[22] Buschmann F., Meunier R., Rohnert H., Sommerlad, P., and
Stal, M. Pattern-Oriented Software Architecture: A System
of Patterns, John Wiley & Sons, (1996), p 31.
[23] Chen, J. An Introduction to Android, Available at:
http://sites.google.com/site/io/an-introduction-to-android,
Last retrieved: 2010-04-09.
[24] Live from Google I/O – Android: Integrate, Replace and
Extend, Available at:
http://mobilementalism.com/2008/05/28/live-from-google-
io-android-integrate-replace-and-extend/, Last retrieved:
2010-04-09.
[25] Sim, N., Turnbull, R. and Walker, M.D. Open devices —
their role in supporting converged services, BT Technology
Journal, 24, 2 (2006), 200-204.
[26] Open Handset Alliance – Overview, Available at:
http://www.openhandsetalliance.com/oha_overview.html,
Last retrieved: 2010-04-09.
[27] Childers, B. Android Everywhere!, Linux Journal, (2009),
186.
[28] Hall, S. P. Anderson, E. Operating Systems for Mobile
Computing. Consortium for Computing Sciences in Colleges,
USA, (2009).
[29] iPhone OS Overview, In iPhone OS Reference Library,
Available at:
ttp://developer.apple.com/iphone/library/referencelibrary/Get
tingStarted/URL_iPhone_OS_Overview/index.html, Last
retrieved: 2010-04-13.