A HYBRID APPROACH TO COMPUTER SCIENCE EDUCATION

A Case Study: Software Engineering at Aristotle University

Andreas Meiszner, Katerina Moustaka

Institute of Educational Technology, The Open University, MK7 6BJ Walton Hall, U.K.

Department of Informatics, Aristotle University, 54124 Thessaloniki, Greece

Ioannis Stamelos

Department of Informatics, Aristotle University, 54124 Thessaloniki, Greece

Keywords: Open source, Software engineering, Open learning environment, Participatory learning, Open participatory

learning ecosystem.

Abstract: Traditionally one characterization of formal education has been that it is ‘closed’, resulting in the fact that

learning spaces with their educational materials, and individual students’ learning processes and outcomes

remain unavailable for the general public. The hybrid approach to Software Engineering piloted at Aristotle

University during the winter semester 2008 / 2009 on the other hand builds upon the way learning and

knowledge creation at the participatory web takes place, in particular within the Free / Libre Open Source

Software (FLOSS) communities. This is to say that on the hand the learning environment used at this course

is open for participation of any individual interested at the subject (inviting in), and on the other hand

Aristotle’s software engineering students are engaging at students driven small scale learning projects, with

each of those learning projects being associated to an open source project (sending out). This combination

of ‘inviting in’ and ‘sending out’ is what we like to call a hybrid approach. One objective of the hybrid

approach is to provide the foundation required for an evolutionary growing learning ecosystem where

learning processes and outcomes have the potential to become learning resources for future students and

therefore connecting content to discourse.

1 INTRODUCTION

There are a number of challenges for formal

education to fully explore the benefits the

participatory web provides for education. With

regards to collaborative learning and knowledge

production the main challenges might relate to the

traditional ‘closed’ and ‘semester based’ structures

of educational systems.

Closed structures on the one hand prevent that

students at one institution could engage and

collaborate at the web in a ‘semi-structured’ way

with peers from fellow universities or the wider

world. This closedness also prevents that the

learning resources of the institution might be

improved by the outside world, or enhanced through

external sources that are brought in by individuals or

through technology.

Semester based structures on the other hand

provide a challenge to establish a learning ecosystem

that would allow for continuous and evolutionary

growth; as well on a community level, including the

full spectrum of participants ranging from newbies

over advanced learners to old foxes, as on a learning

resource level. Such a learning ecosystem would on

the other hand be desirable as it connects learning

resources to learning processes (and related

discourse) or the possibility to establish peer

support, correction, development or even assessment

systems.

A third challenge to education, though not

necessarily related to ‘closed’ or ‘semester based’

structures, is the question how to provide students

with meaningful and motivational learning

opportunities that would allow them to develop their

professional skills within a real world scenario and

impart them as well subject matter skills as also key

Meiszner A., Moustaka K. and Stamelos I. (2009).

A HYBRID APPROACH TO COMPUTER SCIENCE EDUCATION - A Case Study: Software Engineering at Aristotle University.

In Proceedings of the First International Conference on Computer Supported Education, pages 39-46

DOI: 10.5220/0001969000390046

 SciTePress

and soft skills, such as ICT literacy, critical and

analytical thinking skills, project and time

management skills, or presentation, negotiation and

conflict management skills. To respond to this third

challenge web based communities in general, and

FLOSS communities for computer science education

in particular, might be an adequate equivalent to

traditional physical internships, placements or

trainings-on-the-job.

So how to address those challenges that prevents

education within its traditional structures to fully

take advantage of the collaborative learning and

knowledge production opportunities the web and the

FLOSS paradigm provides?

2 FLOSS COMMUNITIES AS AN

EXAMPLE FOR OPEN

PARTICIPATORY LEARNING

ECOSYSTEMS

To deepen our understanding how collaborative

learning and knowledge production takes place at

the web we first reviewed at the EU funded

FLOSSCom project (FLOSSCom, 2008) one of the

likely most mature open participatory learning

ecosystems: the FLOSS communities.

Surprisingly the underlying technology used by

most FLOSS projects is relatively simple, yet

mature, usually including versioning systems,

mailing lists, chats, forums, wikis or similar

knowledge bases. Additionally free web based

services such as Sourceforge provide each FLOSS

project with an initial working and community

environment therefore facilitating the take off of

new projects (Meiszner, 2007).

The way learning takes place in FLOSS is

usually a mixture of more than one approach and

unlike in formal education learning materials are

usually selected by the learner and not the educator.

But more importantly, those learning materials are

commonly generated by the community itself and

also include the code and dialogues between

contributors. Further on students are not acting in

isolation from previous cohorts of students, but the

history of other learners and contributors, and their

remaining availability for follow up contacts,

constitutes a vital element of the learning materials

(Weller & Meiszner, 2008). FLOSS participants also

take on tasks such as knowledge brokering (Sowe et

al., 2006) therefore taking information and

knowledge forward and backward between groups,

communities or even language domains.

From a pedagogical perspective learning in

FLOSS is characterized by self-studying, project-

based learning, problem-based learning, inquiry-

based learning, collaborative learning, reflective

practice or social learning. It is not assumed that

those pedagogies were deliberately set out, but

rather that due to the structure, approach and

governance of FLOSS communities certain

pedagogies have emerged (Glott et al., 2007; Weller

& Meiszner, 2008).

Although institutional education might be seen

today as the prevalent way of learning, self-

education and practical knowledge have their

historical foundations long before the institutional

formal knowledge. Therefore, communities of

common interest like the FLOSS communities, show

how exchange and creation of knowledge can be

supported by the web in a not institutional way.

As described by Glott et al. (2007) one of the

FLOSS characteristic is usually known as

‘openness’ or ‘inclusivity’ of the FLOSS

community. FLOSS communities, like any other

social formation, have established specific cultural

and social patterns and norms that require from

anyone who wants to join a certain degree of

assimilation. Openness and inclusivity does

therefore only mean that those who want to join the

community do not have to pass enrolment

procedures or have to pass formal performance

assessments. Openness also fosters transparent

structures as the FLOSS ecosystem is openly

accessible, including not only code and

documentations, but also communications,

discussions and interactions of any kind, e.g.

through forums, mailing lists or chats sessions.

A second characteristic relates to ‘volunteering’

and ‘volatility’ since FLOSS participants voluntarily

decide which role(s) they want to play or which

responsibilities to take on. As a consequence, roles

and responsibilities (or capacities) of community

members can change over time but also at the very

same time depending on the different contexts. This

results in a very vivid and volatile internal structure

and dynamics of the community (Glott et al., 2007).

A third characteristic is the ‘use of large-scale

networks’ and the way they are established and

maintained. Besides the individual motivational

aspects that must be addressed to attract participants,

and to which we will refer later, FLOSS

communities enable ‘re-experience’, which is a

fundamental mechanism for online learning and

knowledge-building (Hemetsberger & Reinhardt,

2006) and also facilitates new member integration.

Enabling re-experience and the availability of large-

scale networks are also pre-conditions for the

FLOSS volunteering support model.

CSEDU 2009 - International Conference on Computer Supported Education

The fourth characteristics relates to ‘content-

richness’ and ‘specialisation’. FLOSS communities,

though revolving on software development, offer a

range of opportunities to participate that by far

exceed the scope that is closely related to software

(Glott et al., 2007). Content in FLOSS communities

provides users with various types of learning

resources including manuals, tutorials, or wikis, but

also resources that might not be at first recognized as

learning resources like e.g. communications,

discussions or interactions at mailing lists, forums or

chats. One common aspect of the different types of

content is that they are jointly generated by users

and developers and after generation are overall

continuously updated and improved. This however is

not limited to a given FLOSS community, but also

includes the re-use of artifacts that were produced by

other FLOSS communities, or artifacts that are in

general freely available through the web. Those

external sources are usually brought in to the

community by individuals that act as information

and knowledge brokers (Sowe et al., 2006).

A fifth characteristic is the aspect of

‘modularity’, which for the FLOSS case reduces

systemic interdependencies between different files

of the same product, allowing a higher level of task

partitioning and a lower level of explicit

coordination and interaction among programmers.

Modularity might be achieved through a clear

division of labour between the core product and

more ‘external’ features such as modules, add-ons or

plug-ins (Mockus et al., 2000). Within an

educational context modularity might be translated

to organizational aspects of learning, e.g. to allow

participation at a lower entrance barrier, at lower

initial skills, or with less time commitment or more

efficient usage of time available, or to organizational

aspects with regards to modular course design,

including resources created by educators and

learners.

Learning in FLOSS appears to be comparable

with traditional educational settings regarding the

underlying technology and pedagogical approaches

applied, with one of the main differences residing

perhaps on the conceptual and organizational side.

3 POSSIBLE ADOPTION OF

FLOSS APPROACHES IN

EDUCATIONAL SETTINGS

We suggest three different scenarios on the adoption

of FLOSS approaches within educational settings

(Weller & Meiszner 2008; Meiszner et al., 2008),

with each of them having a different level of

complexity and a different degree of benefits:

1. The ‘inside approach’ refers to the practice of

taking the principles found in FLOSS communities

and applying them within the higher education

context. In line with Fischer’s work (2007), this

approach involves mapping the key principles onto

education, including an evolutionary growth of the

course and its environment. This is to say that

current students would build upon the work of

earlier students developing course and content

further year by year, therefore improving content

quality and richness and providing regular feedback.

Such feedback might refer to course structure,

material, processes and tools. The inside approach

thus takes the sort of characteristics and tools found

in FLOSS as its inspiration. The ‘meta-design’

framework and ‘courses as seeds’ process model is

one example for a structured attempt of the inside

approach aimed at supporting self-directed learners

within virtual learning communities by creating

socio-technical environments that support new forms

of collaborative design (Fischer, 2007). Fischer

(2007) talks of users creating socio-technical

environments and has a continuum of participation

ranging from passive consumer to meta-designer.

This mirrors some of the roles of engagement in

FLOSS communities which range from passive

users to core developers.

Within the ‘inside approach’ institutions might

also decide to ‘open up’ their virtual learning

environments to fellow universities or the general

public to view what is going on within the

environment. Within the inside scenario an

institution might even allow those outside groups to

participate and engage at this environment, in the

case doing so, this likely would be a first step

towards a hybrid approach.

A general limitation of the inside approach is

that the outside world remains largely or totally

disconnected, depending on the degree of openness

(e.g. open to view, open to participate, etc.). An

example for a semi-open environment is MIT’s

Open Course Ware project that is partially open for

outside observers, but participation is limited to

formally enrolled students only. Another limitation

relates to ‘community building’ and ‘evolutionary

growth’, since this is per-se limited within a given

institution that only involves the own student

population, and usually even further limited due to

(a) a 100% student turnover per semester / course

and (b) a comparatively small number of potential

community member (formally enrolled students of a

course).

A HYBRID APPROACH TO COMPUTER SCIENCE EDUCATION - A Case Study: Software Engineering at Aristotle

University

The inside approach might be relatively

moderate to implement since the technology should

be already in place at most higher education

institutions, although admittedly modifications very

likely would be necessary. On the down side this

approach still would keep the students of the

institution within this learning environment

preventing their semi-structured engagement and

collaboration within the wider web. It would also

limit the opportunities of ‘best of breed’, as the

wider web might provide better technological

solutions or already established and mature

communities for respective study fields.

2. The ‘outside approach’ at which institutions

would send out their students into already well

established and mature environments to engage at

and collaborate within those communities on pre-

defined tasks. In contrast to the inside approach, the

outside approach might take traditional education as

the starting point by providing theoretical

information and then sends the students ‘outside’ to

find well established communities, such as the

FLOSS ones, to work within those communities and

to apply and deepen their theoretical knowledge.

In particular for the area of software engineering

this approach might be suitable due to the existence

of a large number of mature FLOSS projects and a

myriad of educational resources. This is seen in the

work of the Aristotle University of Thessaloniki in

Greece, where undergraduate students are sent out

into real FLOSS communities as part of their degree

in software engineering. Students are provided with

an initial academic background in principles of

software engineering, testing software and the tools

and approach in FLOSS communities and then are

required to choose and engage with a real project.

This clearly has benefits in computer science as it

gives students real experience of collaborating with

other developers and also of the different types of

roles and work required in software development.

The outside approach, however, is not restricted to

computer programming. It can be realized whenever

there is an external, ‘real’ community that is

operating on FLOSS type principles. The case of

Washington Bothell University (Groom &

Brockhaus, 2007) is a good example for this where

students were required to contribute to actual

Wikipedia articles as part of their assignment work,

thus gaining much of the practical experience of

collaboration and authenticity experienced by the

software programmers at Thessaloniki.

The outside approach might be the least complex

and almost cost neutral; and therefore relatively easy

to implement. The benefits of this approach are that

it responds to the third challenge as mentioned at the

introduction and also would allow for collaborative

learning and knowledge production. However, the

results of this collaborative learning and knowledge

production would remain within this outside

community and therefore likely be lost for future

students. This scenario would also not provide next

year students (newbies) with an easy entrance as no

former learners, nor the resources they created, are

present at the institutional level to facilitate the

newbie entrance.

3. If we view the inside and the outside

approaches as opposite ends of a spectrum, then

there is clearly a range of blended, hybrid

approaches in the middle, which take components of

both elements. Such a ‘hybrid’ approach might be

seen as the best option as it allows a continuous

evaluation (by educators, students and the wider

world) of what ‘the best of both worlds’ is and how

the transferred elements actually suit in their

respective new environments. One of the underlying

assumptions is that using a hybrid approach, as

maybe also partly valid for the inside approach,

could be a response to challenges such as a 100%

student turnover per semester as (a) not all

participating students (and educators) should start at

the same time and (b) free learners outside of formal

education and practitioners are not bound to any

course schedule at all.

Perhaps one such model for this hybrid approach

is that of an open participatory learning ecosystem,

as outlined Brown & Adler (2008). The concept here

is that some of the principles of FLOSS

communities are adopted in education (thus it is an

inside approach), such as collaboration, use of

technologies, or peer production. People learn by

doing, for example by remixing or remashing

content that is viewed by others. However these

activities occur in a broader ecosystem that is open

for everyone combining students, informal learners,

tutors, experts, organizations, etc, and in this manner

it is an outside approach since learners are engaged

in a real global community consisting of a range of

different spaces. Such a hybrid approach likely

would include a number of environments where

students could engage at in a semi-structured way

and where guidance and support is provided through

the use of technologies (e.g. RSS, suggested

contents, etc.) and the use of the human factor (e.g.

knowledge brokers, community support, etc.).

The hybrid approach also has the potential to

open new doors for e.g. (a) new revenue models that

could be based in assessment of learners outside of

formal education against fees and formal recognition

CSEDU 2009 - International Conference on Computer Supported Education

of informally acquired skills, (b) the provision of

niche courses and faster identification of potential

new courses, (c) up to date learning resources and

continuous improvement of processes and products,

or (d) an evolutionary growing community including

the inherent peer support system.

The drawback of the hybrid approach might be

that it probably requires the most drastic overhaul of

higher educational practices and might be the most

complex to implement.

There are a number of cases within formal

education (dePaula, 2001; Groom & Brockhaus,

2007; Wilkoff, 2007; Weller & Meiszner, 2008) that

suggest that the ‘inside approach’ and the ‘outside-

approach’ are viable. Those cases indicate that

FLOSS principles can be successfully leveraged to

educational settings to provide students with similar

learning resources, or allowing them to become

content creators. The hybrid model potentially offers

the highest benefits but remains to be explored.

4 CHALLENGES FOR THE

ADOPTION OF FLOSS

APPROACHES IN

EDUCATIONAL SETTINGS

There are a number of general challenges such as

quality assurance, students’ assessment or cultural

restraints (Schmidt, 2007) that might prevent the

take up of FLOSS approaches within educational

settings. As the outside approach has been already

applied at Aristotle’s Software Engineering course

since the academic year 2005/2006, we would like

instead to address at this section questions that

appeared to us more challenging with regard to the

hybrid approach and to which we might not be able

to respond within the educational framework we are

acting in.

1. The availability of a large number of

(volunteering) participants, which is in the case of

FLOSS communities characterized by volunteering

and volatility, is probably one of the cornerstones of

the efficiency of the FLOSS community as a

learning environment. A crucial question for

transferring FLOSS principles to formal education is

how similar networks can be created within formal

environments, which usually have small classes. On

the other hand, FLOSS community members have

regular contacts to only 1 to 5 other community

members (Glott et al., 2007) and therefore a question

is how to reap similar network effects from small

networks in formal education. Meanwhile the

‘outside approach’ is taking advantage of existing

online communities, the ‘inside approach’ and the

‘hybrid approach’ will need to establish structures,

incentives and motivations to bring together the

different involved stakeholders and to establish such

a community.

2. How to allow re-experience? Within FLOSS

much of the learning processes and outcomes are

made visible and therefore allow future learners to

learn from what others did and to build upon those

experiences – how should this be translated to an

educational setting? A project based approach,

analogue to development processes in FLOSS, might

provide an answer to this as collaboration and

discussions could emerge around those project

works.

3. The motivational aspect: Motivations to

participate at FLOSS are e.g. ‘to learn’, ‘gaining

reputation’ or ‘personal enjoyment’, but also a clear

‘win / win scenario’ between information seeker and

information provider resulting in learning benefits

for both sides (Demaziere, 2006). Those

motivational aspects might be difficult to transfer to

and apply in formal educational settings, where the

main motivation relates to obtaining a formal

degree. While learning in the FLOSS community is

efficient because ‘project managers’ and

‘community managers’ (and many more roles)

voluntarily assume responsibility for organising

work, tasks, content, and communication, in formal

educational settings roles, tasks, and responsibilities

are more pre-determined and rigid (Glott et al.,

2007). And even if allowing for such roles within an

educational setting, what would be the motivation to

assume such roles?

There are a number of possibilities to provide

incentives within formal educational settings such as

rewards for students who voluntarily assume

positions, similar to project or community managers

in FLOSS, or to include into the curricula the

obligation of more experienced students to share

their knowledge with the less experienced. Free

learners outside of formal education might also be

offered a certification of their learning outcomes

against fees, or a virtual credit account that rewards

them for taking on roles such as mentor, facilitator,

moderator or tutor. Those virtual credits than might

be used to pay for assessment and certifications.

With regards to incentives for practitioners to

participate one possibility would be to involve

learner into concrete project works – e.g. to provide

computer science students with the opportunity to

take on some tasks at a respective open source

project. Participants of FLOSS communities are also

A HYBRID APPROACH TO COMPUTER SCIENCE EDUCATION - A Case Study: Software Engineering at Aristotle

University

aware that the skills they learn have a positive value

on the labour market and are able to judge this value

realistically. Precondition for competing with others

that have a comparable formal degree is that

informally attained skills in the FLOSS community

must be provable (Glott et al., 2007). Peer-reviewing

and recognition within the community is very

important in this regard to build up a repute that can

be shown to possible employers. Similar

opportunities, as well for students as for free

learners, therefore might be required within an

educational setting.

But even if addressing all the points above it

might still be a challenge to provide an easy entrance

strategy for own and fellow students, or free learners

outside of formal education. This challenge relates

to questions such as ‘what are learners supposed to

do’ or ‘how to get involved’.

5 META-DESIGN & COURSES AS

SEEDS AS A SUPPORTIVE

DESIGN FRAMEWORK

As a suitable supportive framework for a hybrid

approach we identified Meta-design (Fischer, 2007)

with its underlying courses as seed process model

(dePaula et al., 2001).

Meta-design aims at “defining and creating

socio-technical environments as living entities. It

extends existing design methodologies focused on

the development of a system at design time by

allowing users to become co-designers at use time”

(Fischer, 2007). Meta-design is aimed to support

self-directed learners within virtual learning

communities by creating socio-technical

environments that support new forms of

collaborative design. Meta-design pays tribute to the

fact that future uses and problems of socio-technical

systems can not be totally anticipated by the design

time and must be flexible to changes during use time

and allow an evolution through changed or identified

user needs. Meta-design pays also attribution to the

fact that users might become active participants

within a socio-technical environment that bring in

their ideas and help shaping and forming the

environment and contribute to it. Meta-design is thus

describing relatively precisely what can be observed

within the FLOSS sphere and was therefore seen to

be a suitable supportive framework for the

development of a hybrid learning environment.

6 THE CASE OF SOFTWARE

ENGINEERING AT

ARISTOTLE UNIVERSITY

6.1 Initial Experiences with the Outside

Approach

Since the academic year 2005/2006 the 5

semester

course ‘Introduction in Software Engineering’ at

Aristotle University of Thessaloniki follows what

we have been described as an ‘outside approach’.

The duration of the course is 12, 5 weeks and has

an average student number of 150 with one of the

students’ assignments being to participate at a

FLOSS project or a proprietary software exercise. In

the case selected, the assignment counts for 40% of

the total grade. Also, students can work on their

assignments beyond the 12,5 weeks of the official

lecturing period and submit it at a later time at 3 pre-

defined dates per year – by the end of the course in

February, or alternatively in June and September.

At the year 2005/2006 15 students volunteered

for the FLOSS assignment with the objective of

testing FLOSS and to identify bugs.

In the second academic year of the course

(2006/2007), which had 24 students opting for a

FLOSS assignment, the framework remained the

same with the main difference being that students

now had two options: to test FLOSS or to develop

FLOSS.

At the third academic year (2007/2008) the

framework of conducting the course was modified,

with the only remaining possible assignment option

for students being FLOSS projects, but no

proprietary software exercises. Further to this

students now had three options: to test FLOSS, to

develop FLOSS, or to write a requirement

specification documentation for a FLOSS project

that still had none. For that academic year 55

students have accomplished their assignment by

June 2008.

Motivated by those results, and backed by the

theoretical work of the FLOSSCom project, we

decided to experiment during the semester

2008/2009 with a hybrid approach.

6.2 Design & Trial of the Hybrid

Learning Environment

As part of the FLOSSCom project we developed an

experimental hybrid learning environment

(www.netgeners.net) and run subsequently a small

scale 4 month trial with 10 volunteering students

from Greece and Spain, which were located in 5

CSEDU 2009 - International Conference on Computer Supported Education

different countries and supported on a regular base

by 1 educator and 2 further less regular participating

ones.

This experimental learning environment

provided the same type of tools as identified within

the FLOSS case and tried to take into account

FLOSS particularities such as modularity and

project based work. It was aimed to provide learners

on the one hand with a basic ‘on-board’ set of

communication and collaboration tools (Blog, Chat,

Forum and Wiki) and on the other hand providing a

personal space and a space for personal learning

projects, including rating and commenting systems

as e.g. provided by Amazon.

We than tried to transfer the principle of

modularity and project based work through the

concept of small students driven learning projects

that would allow learners to engage (to a certain

degree) within areas of their personal interest;

individually or together with other learners as a

group work; therefore contributing with their

learning projects to the overall development of the

learning environment and enhancing its richness up

to the point where it might culminates into a very

diverse and rich learning ecosystem.

The concept of project based learning projects

was also seen as a potential bridge between ‘static’

content on the one hand and learning processes and

activities (discourse) on the other hand that might

allow a similar type of ‘re-experience’ as in FLOSS.

Learning projects therefore might allow a FLOSS

type engagement, where content is often taken

forward and backward, contextualized, adapted,

translated, re-mixed, embedded into processes or

feed into new products by individuals. Those

individuals act as knowledge brokers allowing

content to be dynamic and causing it to continuously

change.

This approach did not intend to provide the

learner with a finished set of expert developed

‘static’ content to be consumed, but instead expects

the learner to become an active participant in the

respective study field, to acquire subject matter

skills through practice, and providing the potential

of gaining key and soft skills as a result of their

activities and engagement. An underlying believe is

that for many ‘questions’ or ‘needs’ the answer, or

an approximate to it, is already ‘somewhere out

there at the web’ and therefore, instead of

‘reinventing the wheel’ each time, learners need to

learn how to find, analyse, evaluate and use what

already exists at the web and to incorporate it into

their own work.

Additionally two key aspects of Meta-Design

were considered during the design time:

 “A system should be open to change during

use time and involves all stakeholders in the

design process during design time and use

time“ (Fischer, 2007).

Though the initial core environment has been

largely designed without stakeholders’ participation,

it allowed for stakeholder modifications from day

one of its use time

 “A system should be underdesigned at

design time to allow learners (‘owner of

problems’) to create solutions at use time”

(Fischer, 2007).

This was taken into consideration by allowing

learners to:

 Make use of the communication and

collaboration spaces provided ‘on-board’ or to

use any other space at the web that they felt