BAKER, K. BOWKER, G. Information Ecology - Open System Environment for Data, Memories and Knowing

of 19
12 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Information Ecology: Open System Environment for Data, Memories and Knowing Karen S. Baker, University of California, San Diego, kbaker@ucsd.edu and Geoffrey C. Bowker, Santa Clara University, gbowker@scu.edu Abstract. An information ecology provides a conceptual framework to consider data, the creation of knowledge, and the flow of information within a multidimensional context. This paper, reporting on a one year project to study the heterogeneity of information and its management within the L
Document Share
Document Tags
Document Transcript
  Information Ecology:Open System Environment for Data, Memories and Knowing  Karen S. Baker, University of California, San Diego, kbaker@ucsd.eduand Geoffrey C. Bowker, Santa Clara University, gbowker@scu.edu Abstract. An information ecology provides a conceptual framework to consider data, thecreation of knowledge, and the flow of information within a multidimensional context. Thispaper, reporting on a one year project to study the heterogeneity of information and itsmanagement within the Long Term Ecological Research (LTER) community, presents somemanifestations of traditionally unreported ‘invisible work’ and associated elements of informalknowledge and unarticulated information. We draw from a range of ethnographic materials tounderstand ways in which data-information-knowledge are viewed within the community andconsider some of the non-linear aspects of data-knowledge-information that relate to thedevelopment of a sustained, robust, persistent infrastructure for data collection in environmentalscience research. Taking data as the unit of study, the notion of long-term research and dataholdings leads to consideration of types of memory and of knowledge important for design of cyberinfrastructures. Complexity, ambiguity, and nonlinearity are a part of an informationecology and addressed today by exploring multiple types of knowledge, developing informationsystem vocabularies, and recognizing the need for intermediation. Keywords: memory, infrastructure, information ecology, data management, long-term 1. Introduction This paper develops insights gained from a project that brought together an interdisciplinaryteam to conduct a one year joint study of information management within the Long TermEcological Research (LTER) community. We discuss memory and its relationships to data,information, and knowledge. Memory practices (Bowker, 2005) are at the center of LTER work – the community is aiming to build very long baselines of environmental data, baselines suited tothe life of the ecosystem rather than (as is currently the case) to the lifetime of the researcher. Weanalyze our ethnographic materials to draw out ways that knowledge is held in the LTERcommunity. Such understanding informs information systems’ design and impacts developmentof a robust, persistent infrastructure supporting data work and scientific practices inenvironmental science. We present a conceptual framework for an information ecologyinclusive of data sets and data collectors, information systems and knowledge makers, as well asdigital federations and social networks. The framework is associated organizationally with localdata centers, community learning centers, and global grids, respectively.This work was carried out as part of the NSF Biodiversity and EcoInformatics project entitled‘Designing an Infrastructure of Heterogeneity in Ecosystem Data, Collaborators andOrganizations’. Our interdisciplinary team working collaboratively at the interface of environmental sciences, social sciences, and information sciences (Baker et al., 2002) wascomprised of an LTER information manager, a science and technology studies expert from thefield of communication, an ethnographically trained information systems designer, and the1  LTER community. Ethnographic fieldwork consisted of participant observations, transcribedinterviews, and focused visits to sites, meetings, and workshops. Paper and digital documents aswell as photos were collected across all major roles and categories (site, network, andinformation management). 2. The Ecology of Long Term Databases The database is the cultural and technoscientific object of our times: as rich in its implications ashas been the cinema and the printed book (Manovich, 1999). Charles Babbage (1837) wrote thatthe invention of printing had taken us from being blind creatures of instinct precisely through itsconstitution of a prosthetic memory. However, that prosthetic memory was very limited. Therewas little ability to randomly access books or parts of books without an enormous labor of veryimperfect cataloging, indexing and abstracting. There were few physical copies of informationheld in books, so lumbering mountains would have to journey across Europe to meet their leatherbound Mahomets. With digital databases, we are reconstituting our science, government andarts.And yet the data in databases never stands alone. As Walsh and Ungson (1991) pointed out intheir classic text on organizational memory, there are several different ‘containers’ for memoryin an organization – and these interoperate. Not all information has to be recorded in digital orother archival form. Consider the total institution. Mary Douglas (1986) argues that: “wheneverything is institutionalized, no history or other storage devices are necessary”. If I getprocessed into a prison, I can survive there as just a number (as the Count of Monte Cristodiscovered). There is no need for the institution to hold any information about me other than thatI exist and that I am subject to its regulations for such a time period; there is no need for me toremember anything about my own past, or any sets of skills beyond a fairly simple motor set.Why I am there and who I am just don’t matter to the institution itself – it ‘remembers’ all itneeds to know through the complex set of procedures that it puts into place. Contrast thisextreme example with participant roles in a research community: how much is documented in a job description versus developed in practice; how much is recorded as accomplishments versusaccumulated in experiences which in the case of the LTER community incorporates multi-task roles, cross-site research activities, and interdisciplinary meetings in addition to cooperativeenvironmental field studies. Articulating context introduces one aspect of complexity (Kaplanand Seebeck, 2001; Weick and Sutcliffe, 2001) suggesting that a research community member ora database entry may be wrapped in only a partially articulated context. Multiple perspectivesand criteria introduce a complexity that unfolds into an ambiguity of solutions (Smith Marx,1994; Star and Ruhleder, 1994) and nonlinearity of developments (Wauzzinski, 2001; Yates-Mercer and Bawden, 2002; Solomon, 1997; Spasser, 1997).The replacement of memory by procedures extends to a formal information processingargument that Ashby (1956) made about closed systems of all kinds. He argued that if wecompletely know a system in the present, and we know its rules of change (how a given inputleads to a given output) then we don’t need to bring to mind anything about the past. Memory,he said, is a metaphor needed by a 'handicapped' observer who cannot see a complete system,and the appeal to memory is a substitute for his inability to observe ... . Now no institution isever total, nor is any system totally closed. However, it remains true that there are modes of remembering that have very little to do with consciousness on the one hand or formal recordingkeeping on the other.2  Traces are physical evidences of actions and thoughts. The ecology of memory traces issomething we live on a daily basis. It is rarely theorized as an ecology – more often it is givencompletely differential value: memory held in the head is just not the same sort of thing asmemory held in a file cabinet (Hutchins, 1995). Such a differentiation has some heuristic valuefor us in that it permits several discipline-bound investigations of memory; but it has littlegrounding – it is like the separate ‘pots’ of money (‘rainy day’; ‘college fund’ and so forth) thatpeople create to differentiate their undifferentiable supply of money (it’s just a number, after all).Its limited heuristic value is accompanied by its negative consequence of forcing us into afractured view of our memory work. 3. Long Term Ecological Databases For much contemporary scientific work, data reuse has become a clarion call while also raisingquestions (Zimmerman, 2003). We are currently across the board creating petabits of data – bethey streaming satellite images of our world, probes into space, remote sensors embedded in thewilderness or seismic data echoing from the deep mantle. There is a lot more data beingproduced than there are scientists and techniques to process them. Further, there is aconvergence operating between many sciences around issues of global concern such as thethinning of the ozone layer, climate change, and habitat preservation. The list has been everlengthening since the turn of the nineteenth century (Serres, 1990; NCR, 2001; Ecovisions,2004). In order to answer the questions that the world is posing, scientists require interoperabledatabases – which package time, space, quantity and type in mutually comprehensible ways.They need to be able to share, or at least negotiate protocols between, multiple ontologies.In ecological science, the challenges of unifying time scales, agreeing on spatial units, andclarifying species lists are staggering. Central to long-term ecological studies is the ambition of producing ecological data for the ages (Likens, 1989; Magnuson, 1990). Ecosystems do notdevelop in thirty year chunks (the average career length of an environmental scientist). And yetenvironmental data has in general been collected by a few scientists at most collaboratingtogether for the fixed period of a grant or project. When data collectors retire, typically theirprotocols are not sufficiently well enumerated for future generations to use the data (Bowser,1986). Even when they are well enough preserved, there is the tendency to gather new data withnew tools rather than rework old data. There are at least two prompts for this: one is a feature of professional structures and personal inclinations, which push scientists toward the latesttechnology; another is a fuzzier notion of a learning process, which occurs while scientists aremaking observations and measurements. Facing the scaled-up global task of databasing knowledge about the environment, we confrontdirectly the past of environmentally related sciences. There are huge species lists drawn up atKew Gardens in England and at Harvard in the United States (not to mention lists from non-Anglophone countries). In the nineteenth century, when these lists were first developed, therewas relatively little need to reconcile the respective nomenclatures. Now as our scientific andsocial concerns are becoming global in scope, and as plants themselves are traveling the globe, itis increasingly making a difference whether or not a given designation in Ireland is really or notthe same as one in New Zealand. In the past, such questions have been posed on an ad hoc basis,with taxonomists traveling from botanical garden to botanical garden or receiving typespecimens of plants through the mail. It has been estimated that the rate of synonymy (the sameplant having different names) is of the order of 20% across these lists. Even with these two lists3  reconciled, there remains the work of joining together all of the local lists that draw fromdifferent editions of different standard works in the field (for example, GAP analysis in theUnited States is held to very different State species lists; since State policy for protection of species responds to the local list (Edwards et al., 1995). Even if all the names were agreed upon,it is extremely difficult to conjure information into the right spatiotemporal units. The older databecomes, the more variable the units; however even recent data comes in a staggering variety of forms. Today, technology is enabling the sharing of lists and their organizational structurethrough internet presentations such as the International Catalogue of Life Programme dynamiclist checklist (COL;http://www.sp2000.org/dynamicchecklist.html) and the Global BiodiversityInformation Facility search (GBIF;http://www.gbif.org /portal).The Long-Term Ecological Research (LTER) program consists of twenty-four research teamsof investigators, each team investigating a particular site’s biome in a defined study region(Franklin et al., 1990; Hobbie 2003). Each team works independently to understand the ecologyof their locale; each team also works collectively on cross-site themes and activities. Thisfederation of independent research sites works with a network office and a sense of community.There is a continuity of program funding in renewable six-year cycles, which creates a stableenvironment promoting cooperation and enabling innovative test bed activities. Datamanagement has been a required part of each site’s research program since LTER began in 1980supported by National Science Foundation.Goals of the LTER are to carry out a series of very long-term measurements usingdocumented protocols, to incorporate data management as part of the scientific work itself, andto promote dialogue through shared activities. Here, then, there is the hope of creatingdatabases, which will be useful for very many years, and ultimately across multiple scientificboundaries. Which brings us directly to the question of metadata. The standard response to thedifficulty of creating interoperable databases today is to agree on a set of metadata standards. Ina sense, however, the whole problem just recurses here – since there is a proliferation of metadata standards within environmental science as significant as the proliferation of datastandards themselves. There is little historical evidence that the branching data and metadatastandards can be stopped. This suggests the need to accept that there are very real social,organizational and cognitive machineries of difference which continually fracture standards intolocal versions (Bowker 2005; Boland and Tenkasi, 1995). Rather than see this as a problem tobe overcome solely by another new and better standard, our study of information managementwithin the LTER leads us to propose that a careful analysis of the political and organizationaleconomy of memory practices in interdisciplinary environmental science may lead to thedevelopment of new perspectives in very long term information management 4. LTER Case Study: An Integrative Perspective The field of ecology stresses the links and associations within a system as much as thedifferences and dominions so presents a multifaceted approach to interdependencies of environmental, human, and technological factors, including explorations of principles of self regulation and self correction. The LTER provides an interdisciplinary laboratory withparticipants accumulating shared experiences (Greenland et al., 2003; Robertson et al., 1999;Kinzig et al., 2000). It provides a sheltered forum in which to explore information managementgrounded within a scientific program and to consider the meanings and impacts of interdisciplinarity, data sharing, and technology use on the work of long-term research (Pickett4
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x