Cognitive Artifacts in Complex Work - Ambient Intelligence for Scientific Discovery - page 167

Table 1. Information tasks and artifacts.

Information task Information artifact

1. Locating materials and methods, by searching, locat-
ing, and selecting information from the Internet, primarily
using Google.

“Answer” - In many cases a specific
material or method

2. Seeking journal articles by searching, locating, and se-
lecting from titles and resources listed in the university
library website. This task links to primary resources such
as PubMed or MEDLINE, and specialized services such as
Web of Science.

Article - PDF format to print and
read later

3. Direct searching and selecting of abstracts from primary
search resources (PubMed), linking to PDF versions of de-
sired articles through library-provided article linking.

Set of abstracts or citations to topic
- Print and scan or read at later
time

4. Occasional access of disciplinary websites and tracking
topics through email newsletters.

General research topic or news item

5. Regular scanning of 10-20 journals of specific research
interest, typically online.

Journal, to access articles

6. Continual use of local resources prepared and collected
by the researcher for reading, reference, and immediate use
in research work. Artifacts include printed and annotated
articles, reference lists, abstracts, and electronic versions
of these.

Range of artifacts - Locality serves
rapid access to self-assigned topical
information

Highly Constrained Resource Use. Even though the information resources
in this ecology undergo continual and incremental revision, scientists used only
a very limited set. Across observations, a narrow range of habitual information
tasks were found for all roles studied in the field sample. These six categories
accounted for 91% of cited information tasks, and for nearly all the online tasks
and common resources recorded. Although this finding, while correct, may seem
an apparently impoverished profile of information activity, this task model is
compared with a time-based model of research information practices (Table 2).
A weakness of the task model is its insistence on a single-user, resource-use
model of information use. The “task” model fails to advance understanding of
how information is used in discovery.

4.4 Information Practices in Research
Information research practices are defined as goal-oriented information behaviors
that mediate research activity. Everyday, non-directed interactions are informa-
tion tasks, as shown in Table 1. The difference hinges on the validity, which
we question, of studying user interaction with information resources as a unit
of analysis. Such information tasks are not independent activities, but are con-
tingent on a research purpose. This study established “research project” as the
analysis unit, focused on information use as required by experimental research.
This analysis revealed seven major research information tasks over the life-cycle
of experimental projects. These research tasks are mediated by certain informa-
tion practices as shown in Table 2.

Cognitive Tasks Vary in the Research Life-Cycle. A research project
may require a duration of a year or much longer, depending on how the project
is defined. Discovery occurs as a distributed cognitive process, building upon