Organisational Considerations In The Work Of Software Engineering
Graham Button, Xerox Research Centre Europe, Cambridge
Abstract
The movement towards 'open source' has mainly concerned what are relatively
small undertakings. Yet, some of the problems of large-scale engineering
projects done within complex organisational structures and involving
scores and even hundreds of software engineers may also be addressed
through the open source concept.
Within software engineering organisations, projects
that are managed independently of one another are often addressing similar
problems and might benefit from sharing code with each other, and past
projects may have relevance for current projects. However, the organisational
difficulties that often beset engineering projects may be a barrier
to the idea of open source. For example, software development is done
under strict deadlines and often falls behind schedule, any work not
directly involved in moving the project forward is an unaffordable overhead.
Also, the preservation of code from one project to another is often
jeopardised through lack of documentation and the break up of teams.
Further, developments are often done under a regime of SPI which requires
code validation.
In this paper, we offer a sketch of some of the organisational issues
that surround software and hardware development on middle to large scaled
engineering projects. It is based upon ethnographic studies that we
have carried out on four such projects. As sociologists, it is not our
place to comment upon the appropriateness of open source itself, nor
upon the way in which it could scale up to support the sorts of undertakings
witnessed in our studies. However, with regard to the latter, we outline
some of the organisational considerations that we have witnessed and
the way in which they bear upon the actual doing of the engineering
that may be relevant to a consideration of open source in these environments.
Again, it is not our place to draw conclusions here, but to support
others in doing that by providing evidence and data for their subsequent
deliberations.
|