In 'Lost in the IE', published in the last issue of Ariadne and in subsequent discussion on various blogs ,  there has some thoughtful reflection on the vision of the JISC Information Environment (IE), its architecture and standards, the role of the IE and the role of 'that diagram' . It is clear that the development of work on repositories and services in the UK has benefitted from the IE Architecture diagram but it is also clear that such a model does not (and was not intended to) reflect the reality of the 'messiness' that inevitably surrounds connecting actual repositories and services .
The development, implementation, and support of real services challenges how we have traditionally articulated, represented, and tried to communicate the context of those services. We need abstract visions of an information environment, recommended standards, and models of software architectures (or component software functions) that can inform how we begin to develop local repositories and services. However, we, as a community of managers, librarians, researchers, and developers of technology, also need approaches that help us engage with the complex details of local contexts that shape how and why particular repository implementations succeed or fail. This article, based on work carried out by the Repositories Research Team (RRT) , proposes an ecologically inspired approach to tackle that challenge.
This article will discuss the type of approach we think is missing, and, based on our investigation of ecology as a metaphor, present an overview of an ecologically influenced approach, and discuss an example of its use.
Setting up a repository to manage appropriate copies of, or metadata about, journal articles is perhaps the most understood and established purpose for a repository. Repository software (for example: ePrints, DSpace, Fedora, Intralibrary, Digitool, and Bepress) is designed to accommodate this functional requirement. Standards, such as OAI-PMH and SWORD, have developed with the dissemination of scholarly communication as a crucial use case. Business cases for Open Access have been thoughtfully put forward and numerous suggestions about advocacy to encourage faculty deposit have been explored. Numerous institutions have deployed repositories to support this particular purpose. There is however, no simple, off-the-shelf, recipe to establish a repository for scholarly communications. Selecting and installing repository software is a relatively minor part of the challenge. The growth of institutional repositories for scholarly communications depends on finding the right combination of people, tools, standards, support, networking, and advocacy. Even a cursory review of the stories of successful and unsuccessful repositories highlight many different contextual issues such as: the key role of a particular departmental champion, the importance of good relations with university system managers , the role of funding body mandates to deposit , the benefits of automatically generated publication lists for faculty Web pages , the impact of an Open Access copy on paper citation  and the advocacy value of usage statistics. No one of these factors, or any other for that matter, will ensure the creation of a successful repository for every institution, and to discover which intricate combination might work requires not only trial and error, but also a thorough understanding of the particular complex context of the repository or service.
An essential part of gaining such an understanding of a repository or service in its context is finding a way to articulate and consequently discuss the key dependencies and interactions that shape the repository. This articulation of a complex situation has to be done in a way that can cross skill sets and quickly convey the most important systems, relationships, and other influences that need to be considered for all the stakeholders. However, when we approach such an articulation, we often pick an analytical modelling tool, such as UML (Unified Modelling Language)  or BPMN (Business Process Modelling Notation) , to represent the situation with that tool, and use that as the basis of our presentation and analysis. As useful as this can be, it is –if used on its own – also an inadequate and potentially flawed approach. This is not as a result of an inherent difficulty with the tools but rather because we are not adequately addressing the true nature of a complex situation.
Ritchey (2002), drawing on Ackoff (1974) and Pidd (1996), summarises a view that there are three types of complex situations in what he terms a socio-technical system; these are: messes, problems, and puzzles.
Ritchey captures implications of the above distinction between 'mess', 'problem', and 'puzzle' in the following quotation from Pidd,
One of the greatest mistakes that can be made when dealing with a mess is to carve off part of the mess, treat it as a problem and then solve it as a puzzle -- ignoring its links with other aspects of the mess.
Repositories and services often exist in this sort of mess. Not as a result of any failing or sloppiness on the part of the managers or developers, but because as organisational, technical and cultural entities - like many other services and systems - repositories exist in the midst of an extremely complex set of interactions and influences (only a small percentage of which are technical). As a community we run the risk of reducing this complexity to whichever form of transaction that fits with our chosen problem-solving tool. If we do this, we sidestep some of the complexity of the mess, view only one dimension of the problem, and then treat it as a puzzle. As a result, the proposed solutions may neglect the context they are intended for and consequently struggle to succeed because the scope of the tools was too narrow, and critical factors were overlooked. We need approaches to support discussing the 'mess', and that help us to draw out the problems that need to be considered. To this end, the Repositories Research Team (RRT) has examined the possible benefits of using ecology as a metaphor to shape how we, as a community, articulate complex situations.
Ecology – 'the branch of biology dealing with the relations and interactions between organisms and their environment, including other organisms' – examines highly complex processes and interactions as they affect a particular organism being studied . Although a branch of biology, the interactions it considers are not restricted to the biological; other types of interaction, such as chemical, physical, meteorological, geological, may play a part. For example, the survival of a herd of deer in a particular area will depend on multitude of factors including the availability of food, the presence of predators, the climate, hunting restrictions, and building developments. Investigating the details of such an ecosystem and assessing its sustainability could be a detailed and lengthy process, but a basic understanding of it is also possible without specialist knowledge.
If we consider the allotments pictured below as an example of an ecosystem, we intuitively know that the success or failure of one of the plots, such as one on the middle right, will depend on a complex set of interactions. The interactions between the gardener, the location of the plot, the chosen crops, weeds, the neighbouring gardeners and any wildlife, form a small ecosystem. The whole ecosystem is affected by differing environmental factors ranged from the weather and the location of the plot relative to the shed, to how often the gardener can visit the plot, local byelaws and regulations.
Perhaps because, at least on some levels, people seem to have an intuitive understanding of ecology, it is often used as a metaphor for complex situations. It has been used before in the field of library and information science, by Nardi and O'Day , and by Davenport , and in connection to digital libraries and repositories by Blinco and Maclean . Our investigation explored the relevance of the metaphor for repositories, services, and their interactions, to see what concepts and approaches might be of use in helping articulate specifics of complex dependencies and supplement existing analytical tools and skills. An ecological approach to repository and service interactions , the introductory report, we produced on this topic, goes into more detail about how these related concepts might be applied, the following section briefly presents some ideas that may be of use.
We suggest that the metaphor of an ecosystem provides the following useful concepts that can help articulate complex situations:
An ecological metaphor also provides us with some useful concepts that can inform how we approach modelling and how to assess the interactions being represented. These are concepts that provide an easily grasped way to present and analyse a set of complex interactions:
Models using ecology as a metaphor are not as restricted as traditional problem-solving tools and allow the key parts of a complex situation to be expressed. The intuitive nature of the ecological metaphor may be more accessible to a larger group of people. The development and availability of such a model facilitates discussion and allows non-specialists to grasp the key features of a setting. This may lead to shaping a complex mess into discrete problems that can be tackled with other tools.
We recently presented a poster at DC2008 about an academic's dissemination of a presentation, specifically looking at what happened to his presentations and associated metadata . A brief examination of a couple of its key features can illustrate one outcome of this type of ecological modelling.
The poster presented the ecosystem of a researcher 'Phil Robertson' trying to disseminate a presentation. Phil created a workshop presentation based on a recently published paper. After the workshop he sought to make the presentation more widely available. The repository at his institution had accepted a preprint of the paper, but it was not interested in the presentation. He uploaded a copy of the presentation onto SlideShare and blogged about both the paper and presentation. The presentation was then discovered by a number of researchers. The entities included: Phil, the institutional repository and its manager, SlideShare, flickr, the CETIS blog, two researchers, the journal publishing process, and the abstracting and indexing process. Some of the interactions were: manage account or profile, cite, deposit, and discover. Key environmental factors were: the collection policies and rights policies associated with the different tools, the network of the academic discipline, and social networks around Web2.0 tools. The poster primarily took a resource-tracking approach to study and asked questions about the metadata associated with the various forms of the work and their different points of access.
Two key issues emerged from being able to have an overview of this ecosystem that could articulate technical, cultural, and human context of this researcher's efforts to disseminate his work. These were the differences between the author identities and the challenge of preservation.
The different tools that academics use to disseminate their work all have different ways of creating and maintaining that author's identity. In this instance, the identity associated with the paper carries a different organisational affiliation to that of the identity associated with the dissemination of the presentation and the blog. The paper carries an institutional affiliation requested from the author as part of the paper; while the repository carries a departmental affiliation created as part of a profile (with the institution implied). On the other hand, the blog, hosted by an organisation Phil belongs to, carries organisational branding and affiliation; whereas the academic's identity on SlideShare carries no affiliation or other identification beyond what he added to his profile when it was last updated.
If we consider questions about the access to the presentation in the longer term, it is clear that, outside any personal archiving, long-term access to Phil Robertson's presentation is dependent on the existence of SlideShare, and its continuation of a business model supporting free access to the content it stores. Discovery of the presentation in the longer term may also be a challenge. If the link between the blog and SlideShare were broken, SlideShare only supports very basic metadata (a title, tags, and an association with an optional author profile). Practitioners seeking to preserve presentations as part of their work will encounter considerable difficulty unless they actually know the author's SlideShare identity.
We found that this ecological model provided an easy way to identify, highlight, and discuss key issues about managing presentations. It could also be used to assess quickly the impact of a desktop deposit tool on the current dissemination workflows, and to identify the technical, practical and administrative issues such a change would raise.
In this article we have highlighted the need for approaches to supplement the current use of tools and models within the digital library, repository and service community. We then introduced our exploration of ecology as a possible alternative metaphor for work that has been carried out in connection with it. We introduced an example of a developed ecology and illustrated its usefulness through a few of the issues it had raised. The next stage of development of our investigation of repository ecology is the release of the final version of the introductory report and the completion of three more case studies of using this approach.
It will, we hope, be clear that the ecologically influenced perspective we are suggesting is intended to support discussion and reflection on how we implement and use systems. In providing a way to articulate the context of a service more fully, we hope that this will lead to richer discussion about why repositories succeed or fail. We hope it offers repository managers a way to model the context of the repository that does not rely on knowledge of a particular modelling tool and which can engage a variety of stakeholders (whether librarians, developers, academics, or administrators).