2012. Vol.3, No.2, 269-274
Published Online April 2012 in SciRes (http://www.SciRP.org/journal/ce) http://dx.doi.org/10.4236/ce.2012.32042
Copyright © 2012 SciRe s . 269
Learning the Business
Mike Holcombe, Marian Gheorghe
Department of Computer Science, University of Sheffield, Sheffield, UK
Received November 7th, 2011; revised December 8th, 2011; accepted December 28th, 2011
Developing software is a highly creative process. This paper describes a novel approach to teaching soft-
ware engineering which involves university students working in partnership with external clients from
business, charities and the public sector building solutions to their business and other problems. The paper
describes the basic principles behind these activities and focuses on the experiences of teaching advanced
students through the medium of a commercial software development company specifically set up to be
run by the students as part of their degree course. The evidence from student and employer feedback
demonstrates that this approach has been highly successful for the past 20 years or so but despite this it
has rarely, if ever, been replicated elsewhere.
Keywords: Software Engineering; Commercial Clients; Projects; Enterprise
Genesys Solutions is a software house with a successful list
of commercial clients developed over the past 16 years. It is
entirely run by senior students, 4th year undergraduates and
MSc students, in the Department of Computer Science as part
of their computer science degree. Students are responsible for
marketing and sales, all software development and the man-
agement of the company’s computer infrastructure. This ex-
perience enables them to undertake the complete range of ac-
tivities that any successful software company carries out from
consultancy services; contract negotiation; development of high
quality solutions for their clients; customer relations; software
maintenance and the setting of company strategy and business
planning. All of these activities require a challenging mix of
creative, logical, quality-oriented and entrepreneurial skills.
Designing software to solve real problems is a difficult crea-
tive activity. In general, software engineering follows an ap-
proach similar to the Creative Problem Solving Process (CPS),
[Osborne & Parnes (1950)]. CPS is a structured method for
generating novel and useful solutions to problems. The first
phase is Exploring the challenge, then the Generate ideas stage
follows and finally the implementation and evaluation of the
solution—in this case a software application. This has to be
done in the highly constrained and often unforgiving environ-
ment of a programming language. This software engineering
involves both creativity and discipline with quality assurance a
vital underpinning theme.
However, the traditional CPS approach doesn’t work in
situations where the problem is ill-defined. This is usually the
case for most software development projects. The process must
be refined and made much more iterative, culminating in the
agile process that we now use. Initially the developers—stu-
dents in this case—talk to the client about their business or
organization and their objectives. This discussion then starts to
focus on what possible solutions might look like. This often
includes some rapid prototypes and screen shots to enable both
sides to understand then issues better. A set of tests are pre-
pared to enable both sides to see if what is being done works.
This emphasis on how detailed evaluation will be done at the
start of the project is very important. Then a development stage
could be carried out and further reviews and exploration carried
out, thus repeating the process. This might lead to significant
changes to the project and is a key aspect of agile development.
The client’s business environment might change and this could
lead to changed requirements and the decision to throw away
some of what has been developed. Such an approach brings
with it very strong demands and challenges for students but is a
good experience in terms of demonstrating that the real world is
messy, uncertain and volatile—something that they will have to
deal with in later life.
This emphasizes that creativity, in reality, often exists in a
dynamic environment and adaptability and quality assurance
are key issues to be mastered.
For many years the Department of Computer Science at the
University of Sheffield has put great emphasis on the develop-
ment of team skills in realistic settings—in particular the sec-
ond year Software Hut module involves teams of students
working with external business clients developing a software
solution for their business or organization [Kalra et al. (2005)].
This is carefully managed with the aim of developing a profes-
sional attitude in the students and using a management frame-
work and quality development tools that usually ensures suc-
cess as well as a powerful learning experience for the students.
This has been running since 1987 and was highly praised by the
UK Government’s Teaching Quality Assessment visit in 1994.
The overall undergraduate programmme places group projects
at the core of the curriculum. Thus students, on their first day at
the University, are placed into teams and spend one sixth of
their first year going through the main stages of the software
engineering lifecycle. Thus they learn about business analysis
and requirements capture, write a short “White Paper”, develop
a requirements document and an acceptance test report for a
“pretend” application over the first few weeks. The teams then
exchange these documents and receive the documentation about
a different application from another team ready for the next
stage, which is the development of a design document. This
M. HOLCOMBE ET AL.
takes place in Semester 2 and is followed by the implementa-
tion stage—again swapping applications to a 3rd example. Fi-
nally they receive the documents and code for their original
system and evaluate it, writing a report on this stage. This ex-
perience is intended to underline the importance of coherent,
informative and concise documentation and provides them with
some wider-ranging experiences of building a piece of software
from concept to delivery. This is the foundation for the 2nd year
Software Hut experience where they are working with a real
external business client.
In year 3 the students carry out an individual research project
alongside their advanced technical courses.
Following this, a grant was awarded (under the Fund for the
Development of Teaching and Learning) to develop the ideas
further. This coincided with the UK Government introducing
the SARTOR (Standards and Routes TO Registration) for the
engineering profession that required the introduction of a 4th
year of study for engineering accreditation. It was decided to
introduce a radical new approach into the 4th year that built on
the Software Hut experience.
A company was set up within the University which would
carry out commercial software development activities and
which would be largely run by the students as part of their 4th
year studies. This article is a record of some of the activities,
successes and lessons learned during the past 16 yea rs.
The Early Days of the Company
Initially, the numbers of students taking the 4th year were a
minority – less than a dozen, so the company started small and
this provided the academics with an opportunity to learn
quickly and without too much embarrassment! The first issue
was to find some premises in which to house the company and
some computing infrastructure. This was achieved by the use of
a small laboratory in the Department and money from the
FDTL grant. At the beginning these machines were managed by
the Department’s technical support and the company concen-
trated on getting business and developing both a set of com-
pany processes and a good reputation. The students were given
a great deal of responsibility for these issues and this worked
out very well as they were both motivated and sensible about
what was introduced. The students negotiated contracts and set
the price of projects with advice from us. The general policy
was to charge at a rate that the clients could afford rather than
at commercial rates. This allowed us to develop a niche market
that did not compete with established software companies.
Genesys has provided many small start-up companies, charities
and others with an opportunity to have bespoke software cre-
ated for their businesses and organizations that they would
otherwise never have been able to afford. This in no respect
made the intellectual challenge of the projects different from a
full commercial one.
In fact, the decision to give the students a lot of responsibi-
lity for running the business has proved to be a very successful
one. Initially the main focus of activity was on producing fairly
standard software systems suc h as databases and planning tools
for a variety of organizations—small business and charities and
public sector departments with the health service providing
many such projects. These projects used a fairly standard soft-
ware development approach with the students working closely
with their clients and managing the projects. As their supervi-
sors the academic staff met the teams regularly and provided
support, training and advice. The company had regular board
meetings that were chaired by a student and the emphasis of the
meetings were to provide a means of sharing information about
the projects and the technologies used, deciding on which fu-
ture contracts to accept and planning out the company strategy.
This latter included: defining the company profile, corporate
image, web site as well as the internal processes, quality assur-
ance, standard tools and templates. The students had many
good ideas and usually made best use of the latest technology.
Business was fairly easy to come by—it was mainly through
word of mouth or contacts inside the University and research
collaborators in other universities. The new course was becom-
ing more popular and reached a point when there were 50 stu-
dents in the company. They spent one sixth of their time doing
the module and it soon became clear that this was not enough—
both the students wanted to spend more time in the company
and the business was there to justify it. One other factor was
that there was some risk that the enthusiasm for the company
meant that some students spent more time on this work with the
result that their other modules suffered. We thus agreed to dou-
ble the time for the course—this meant that the students were
expected to spend 15 hours per week—one third of their study
time. This has remained in place and has been a positive de-
By this time we had our own premises away from the De-
partment and our network was completely independent from the
University’s. This meant that some of the students had to be-
come fully-fledged systems administrators running a network
of up to 30 PCs with full file store, software infrastructure,
internet connection, a highly secure system and full 24/7
back-up. I have always been impressed at the professionalism
and technical skills that these students display—in many cases
the systems they run are better—more modern, more reliable
and generally provide a better service than some offered by
full-time professional services!
At the end of their studies the company board meeting, in-
cluding the students, looked at the balance sheet and the stu-
dents decided what to do with profits—typically this involved
holding back some for reinvesting in the company infrastruc-
ture with the rest being split amongst the students as a cash
payment. Of course, the students were also getting academic
credits towards their degrees.
We have also had a long-ter m relationship with IBM ( Hur sl ey
Park) who have provided three experienced developers as
mentors for many years. These mentors come up every term
and provide advice, training courses, as well as “adopting”
some of the teams having telephone conference calls and e-mail
discussions at other times. This is highly appreciated and the
benefits are in both directions.
Becoming a Fully Constituted Company—The
Next Stage in the Company’s History
The students were running the company during term time but
in vacations there was a gap. On occasions there were opportu-
nities for business over the vacations that we couldn’t take ex-
cept at the cost of employing students after they had graduated—
this we did a few times but the administrative load was rather
heavy. There was also the issue of maintenance which did arise,
usually it was for an extension to a client’s system but some-
times it was necessary to fix some problem. With an “empty”
company over summer this was a potential issue.
Copyright © 2012 SciRe s .
M. HOLCOMBE ET AL.
Copyright © 2012 SciRe s . 271
The decision was taken by the University to spin out the
company with a slightly different role. Some funds were avail-
able from the European Union to support this process. The new
company—called epiGenesys—was duly constituted and regis-
tered with the Government. A legally constituted board was
established with the University being the sole shareholder.
A major role for the new company was to support the Uni-
versity’s exploitation of its research Intellectual Property th-
rough the transformation of experimental software developed
by research projects into commercial quality products. In order
to achieve this some permanent employees were appointed
from amongst the best of the Department’s students and gradu-
ates. Currently there are eight staff. They have a dual role—
managing the epiGenesys project list and mentoring the stu-
dents in the Genesys company—Genesys is regarded as a sub-
department of epiGenesys. To support the permanent staff the
company faces a challenging business revenue target which has
concentrated all of our minds.
Early Adoption of Agile Development
It became clear around the turn of the century that the “tradi-
tional” design-led approach we were taking in the company was
not as appropriate as it might have been. Many of our clients
were unsure of their precise system requirements and the proc-
ess of business analysis and requirements capture was a rather
dynamic one. The emergence of Extreme Programming (XP)
and the work by Kent Beck [Beck (1999)] came at just the right
time—we adopted XP in 2000. This has been a great success
and we have increased our efficiency and quality substantially
through this. We undertook a large funded research project into
the comparison of the agile development approach compared
with the traditional software development one using the exten-
sive project data we collected, including a large number of
comparative experiments where different student teams used
the different methods on the same projects. This produced a
clear advantage in terms of quality of delivery for the agile
approach. It is also greatly enjoyed and preferred by students
and this was established through measuring well being during
the projects as well as other samples of student attitudes.
As we explored the detailed processes of XP we developed
our own software infrastructure and management approach.
This is the subject of the recent textbook [Holcombe (2008)]
which tries to look at software engineering in a broader than
usual perspective with an emphasis on “people” issues.
The epiGenesys Way
The basic software development approach of the company
involves a rigorous version of eXtreme Programming (XP). As
far as possible students work in pairs on the projects. The pairs
change around regularly. Although many students were scepti-
cal about this initially most recognize the great benefits that it
brings once they get involved in a challenging project. The
initial phase involves meetings with the clients and the analysis
of their business needs and context. This is done in a number of
ways—interviews, questionnaires, role playing etc. and the
outcome within a week or two is an outline requirements de-
scription and a set of stories. These represent small elements of
functionality that can be implemented in a week (15 hours of
pair working). A tool, StoriPost, has been developed to provide
support for the management of the stories. A screenshot of this
tool is in Figure 1.
Each story is described according to the template:
As each story is prioritised, worked on, tested and completed
the story “postit” is moved right to the next column. The details
of each story can be opened up as in Figure 2.
It is important for a company to develop its own tools and
processes to both improve productivity but also to build a
company culture. StoriPost is one example of this.
A dynamic summary of the stories and solution is generated
as a lightweight requirements document since many clients
require such a statement.
Depending on the type of application frameworks such as
M. HOLCOMBE ET AL.
Ruby on Rails, and in-house systems developed in epiGenesys
such as PHP-trax are used. Use of Subversion (svn) a version
control system is mandated. Continuous integration is sup-
ported using cruise control.
Testing is taken very seriously and we try to use test-first
techniques, which means writing the tests before coding. The
acceptance criteria specified in the stories determines the unit
tests. Testing tools such as phpUnit are used for unit testing the
stories. An architectural design notation—XXM (extreme X-
machines, Holcombe (2008) is a popular way to illustrate how
the stories are integrated through a user interface. An example
is shown in Figure 3.
Systems test—particularly those systems with a web front
end—use Celerity and in-house feature converters using story
and XXM information to automate the filling of web forms and
actions in testing.
Other aspects of the software development are covered by
appropriate tools: processes are handled by Basecamp, other
types of testing by Cucumber, training Genesys students at the
beginning of their programme with in-house built tools.
A brief selection of projects includes:
1) Bummitt—a social networking web site where teams of
students competing in a charity dash to see how far they can get
in continental Europe on £15 and which shows maps of pro-
gress based on receiving text messages from the teams;
2) Iceberg tracking software for shipping based on some pro-
totype software developed by a research team;
3) A web site for children suffering form Cystic Fibrosis that
helps them, through the use of a game, to calculate their medi-
cation based on what they have eaten that day;
4) A video game about the 100 years war in France in the
middle ages that features manuscripts, weapons and other arti-
facts being showcased in a museum exhibition;
5) A system that allows the audience at a meeting to vote on
questions posed by the presenter using the Bluetooth capabili-
ties of their mobile phones.
There are many other examples.
It is quite clear that this activity greatly increases all the stu-
dents’ motivation and develops their skills in many areas to an
extremely high level. They are real professionals when they
finish and can walk straight into an industrial position and be
productive immediately. Very few graduates are in this position.
All the graduates from this course are highly sought after by
companies large and small.
The popularity of the course is extremely high with well over
90% of students confirming that this was the best thing in their
degree programme and stating how much they enjoyed the
experience even though it was very challenging.
All our evidence indicates that the course meets the needs of
both students and employers. Not only that but some of the
students go on into academic research and make that a great
success as well—they are highly skilled, highly motivated and
very well organized—key attributes for a successful researcher
The Benefits Are for Everyone—Win, Win,
The benefits for students are clear from the above, but there
are also benefits for the academics involved in that it forces us
to address the realities of real software development in a com-
mercial context and demonstrates that much academic thinking
in software development is based on myths. Our research has
benefited because we can carry out empirical research into
development methodologies, study team related issues and base
this work on real commercial development projects. Much of
this experience has been incorporated in [Beck (1999)].
Some universities encourage or mandate students to spend
some time on an industrial placement. These can be very re-
warding experiences during which the students can gain many
new skills. However, such placements can vary greatly in the
quality of the experience for the student. In some cases they
provide an excellent vehicle for improving their knowledge and
professionalism. However, in some cases the students do not
gain as much as they could. I have heard of many cases where
the students have been given mundane tasks and little opportu-
nity to engage with an important project. They rarely meet with
clients, for example. They also rarely have any strategic input
into company policy or the opportunity to take major responsi-
bility for delivering a successful product. All of these aspects
form part of the Genesys experience. One of the members of
the Department’s Industrial Liaison Board once commented
that there were two types of software engineers working in
industry—the foot soldiers (programmers) and the leaders—
project leaders, business leaders etc. One of our targets is to
educate for the latter posts as well as the former.
The local community also benefits, there are many small
start-up companies that have been able to progress using soft-
ware developed by the students—this includes charities and
public sector organizations as well.
The University benefits because it is seen to be doing some-
thing useful for the community and it is attracting a lot of good
students—unlike many other universities Computer Science at
Sheffield is booming in terms of student recruitment.
Copyright © 2012 SciRe s .
M. HOLCOMBE ET AL.
An XXM from a project— s e e (Holcombe, 2008).
The United Kingdom Government collects various perform-
ance statistics about university courses and these inevitably fine
their way into league tables published by the media. One aspect
is graduate employment and this is measured by the Graduate
Destination survey—all graduates are contacted 6 months after
graduation and recorded as 1) being in graduate level employ-
ment; 2) undertaking further study or 3) not in either of these.
Our students have consistently been reported with an unem-
ployment rate of close to 0% for the past few years placing the
Department at the top of the list.
Another major survey is the National Student Survey. Here
students are contacted on behalf of the Government by a survey
firm and asked about their satisfaction with various aspects of
their degree course. Again, every course in every university in
the United Kingdom is surveyed and in the 2009 survey the
University of Sheffield was top for overall course satisfaction
in Computer Science gaining a 100% rating. There are 140
institutions offering computer science and related degrees in the
United Kingdom and none others received this rating.
There is thus strong evidence that out approach to practical
project work is a major contributor to these results.
A question we are often asked by business executives, politi-
cians etc. is—why doesn’t everyone else do something similar?
There are a number of reasons for this—it is difficult to start up
something as innovative and unprecedented as this because of
the risk aversion and bureaucratic constraints that university
teaching experiences currently. Few academics have the ex-
perience to try something like this on, it seems. All of these
obstacles applied to us when we started but, perhaps though
naivety or just sheer bloody-mindedness we just got on with it
and made it work. There were lots of problems on the way and
we learnt rapidly but it was certainly worth it in the end.
Enterprise can be taught successfully in universities but the
way it has to be done is radically different from our normal
educational processes. Educationalists have to be bold and
adopt a similar real life oriented strategy. It takes courage but
the learning experience for academics amazing.
However, there are responsibilities for employers here, also.
Sometimes our enterprising graduates join companies with
great expectations only to be disappointed when the companies
Copyright © 2012 SciRe s . 273
M. HOLCOMBE ET AL.
do not allow them to express their skills and they become
bogged down in the bureaucratic culture of the company. In
such cases, and they have happened, the students soon leave
and join smaller, more dynamic business. There is thus a two
way responsibility, we can produce more enterprising graduates
but some companies need to change to get the best out of them.
Osborne, A., & Parnes S. J. (1950). Creative problem solving: Re-
sources for CPS practitioners. Charlotte, NC: OmniSkills, LLC.
Beck, K. (1999). Extreme programming explained: Embrace change.
New York: Addison-Wesley.
Holcombe, M. (2008). Running an agile software development project.
New York: Wiley. doi:10.1002/9780470385883
Kalra, B., Thomson, C., & Holcombe M., (2005) The Software Hut—A
student experience of eXtreme Programming with real commercial
clients. Extreme Programming and Agile Processes in Software En-
gineering Proceedings, 3556, 323-324.
Copyright © 2012 SciRe s .