Awareness is the coordination mechanism, and it is also the channel for surveillance

A shared workspace coordinates by showing its occupants to one another. Dourish and Bellotti (1992) gave the field the observation the third of ten units starts from: the information holding a group's work together is usually generated passively, as a by-product of activity in the shared object. That by-product has a second audience, since the trace letting a colleague avoid duplicating your work is also a record of your working day. Awareness is therefore the coordination mechanism of shared workspaces, and also the channel through which they are monitored. The design question is not how much to show but how to distribute it: symmetry, decay and aggregation. The next unit asks what distance costs.

The signal that does the coordinating

Dourish and Bellotti (1992) separated active awareness, which a system produces by asking people to report what they are doing, from passive awareness, which the shared workspace produces as a side effect of work. Active mechanisms, they argued, penalise the supplier, presume the supplier's sense of relevance matches the recipient's, and let the sender control delivery, so the recipient cannot browse it when it becomes useful. Their case was ShrEdit, used by groups of three designers in separate locations with ninety minutes to design an unstaffed post office. It showed others' edit actions but not their cursors, and the designers complained about the missing cursors.

Coordination without asking, in working settings

Heath and Luff (1992) studied the Line Control Rooms of London Underground, a complex multimedia environment in transition, and explicated the tacit practices through which personnel communicated and coordinated many tasks. Tang (1991) videotaped small groups at a shared drawing surface and found hand gestures carried significant information.

Bly (1988) compared two-person design sessions face to face, over an audio and video link, and by telephone (Bly, 1988). Many of the same drawing activities appeared in all three settings, even though they did not share the same surface, so drawing carried information for the designer and not only for a reader. The finding bounds the mechanism as well as supporting it: the study was small and preliminary, and a surface that cannot show another person's marks still carried much of the work.

A framework made awareness analysable, and the field then contested the term

Gutwin and Greenberg (2002) built a descriptive framework for workspace awareness in real-time distributed groupware, in three parts. It names the elements of knowledge that make up awareness, the perceptual mechanisms that maintain it, and the ways people use it. Who holds elements about whom, how long a mechanism retains them and whether a use receives events or a summary become choices a designer sets. This unit calls them symmetry, decay and aggregation.

The concept is contested from inside the field. Schmidt (2002) introduced the journal's special issue and reconstructed the term's history. Awareness entered CSCW as a placeholder for elusive practices of taking heed of what is happening. The media space programme expected open audio and video links to reproduce face-to-face cooperation, and that expectation never materialised. He argues the term does not denote a set of related practices, so a requirement written as support awareness specifies nothing a mechanism can be built against. Schmidt's own formulation is friendlier to this unit's thesis. Monitoring and displaying are mutually dependent.

Awareness can be set as a quantity, and it reached a population

Benford and Fahlen (1993) made awareness a quantity. Their spatial model defines object aura, nimbus, focus and adapters to control mutual levels of awareness, so what one person perceives of another can be set. Cadiz and colleagues (2002) show the question is not confined to prototypes: Sideshow, a peripheral awareness interface, was installed by 13,000 employees within a year and used regularly by 7,000. The same data carry the cost: lost screen space was the primary reason for stopping for 8 per cent of respondents and a secondary reason for 15 per cent. Awareness signals are mostly visual, so a user with a screen reader is shown less than the interface assumes.

The same signal is the monitoring channel

Hudson and Smith (1996) stated the structural problem. They describe a dual tradeoff. The first is between privacy and awareness: the more information about oneself leaves the work area, the greater the potential both for colleagues' awareness and for intrusion. The second is between awareness and disturbance: the more a person receives about colleagues, the greater the chance it disrupts their own work. They call the tradeoff fundamental. Three variables locate a system on it.

1. Symmetry: who receives information about whom. Schmidt's (2002) reciprocity is why this comes first: the benefit and the risk are one signal, so reducing the risk means changing its distribution.
2. Decay: whether a trace persists. A fading signal supports coordination at the moment of work; a retained or searchable one becomes a record usable for other purposes.
3. Aggregation: whether a recipient sees events or a summary. A summary keeps the indication of activity and withholds the particulars that make a working day legible.

Sum, Shi and Fox (2025) analysed Reddit posts and worker interviews. They found layered surveillance met with little transparency or autonomy, answered by resistance that stayed individual.

Designs choose what to transmit, and the attention cost can be evaluated

Erickson and Kellogg (2000) answered with social translucence. Visibility, awareness and accountability then let people structure their interactions, as their deployed system Babble does. Ackerman and Starr (1995) answered with aggregation: indicators of social activity built on a CSCW toolkit, because people attend to a system when they know there is activity on it. Prinz (1999) answered with infrastructure: NESSIE is application independent, with configurable indicators and a server supporting access control, reciprocity and subscriptions on shared interest profiles.

Carroll and colleagues (2003) press on the disturbance half. Analysing breakdowns in a collaborative learning system, they argue that notification supports awareness of presence and actions but not of persistent, complex activities, and propose activity awareness, which adds planning and coordination. Pinelle, Gutwin and Greenberg (2003) make the coordination claim testable. Their Collaboration Usability Analysis models the teamwork in a group task rather than the taskwork, and it grounds actions in primitives they call the mechanics of collaboration. If awareness is the coordination mechanism, an evaluation modelling the shared task should find problems the interface alone does not show.

Contested question: can the coordination benefit be separated from the monitoring risk?

The first of three positions says the benefit and the risk cannot be separated. Hudson and Smith (1996) call the privacy and disturbance tradeoffs fundamental, so every increment of awareness carries a monitoring or attention cost, and Schmidt's (2002) reciprocity points the same way. The second says they can be separated by design: Erickson and Kellogg (2000) with translucence rather than transparency, Ackerman and Starr (1995) with indicators that summarise, Prinz (1999) with configurable access control. The third concerns the software's role: Heath and Luff (1992) found coordination running through a shared environment and tacit practice, so in rich settings the mechanism is the room, while Dourish (2001) denies a system holds context at all.

What would settle the first two is a deployment measuring both sides. Hold the work constant and vary symmetry, decay and aggregation in one system. Measure coordination outcomes such as duplicated work alongside monitoring outcomes, such as whether traces are consulted by those with authority over the people they display. Prinz's (1999) infrastructure is the nearest existing instrument, since reciprocity and access control are configurable in it, though it was not run as a test of the tradeoff. The third is settled by method, a comparison varying the physical arrangement while the software is held constant.

Hudson and Smith (1996) argue from theory rather than report a measurement, and Schmidt (2002) argues from a failed programme. The separating reading rests on three designed systems, none a comparison of the tradeoff (Erickson and Kellogg, 2000; Ackerman and Starr, 1995; Prinz, 1999). One reports numbers, since Sideshow reached 13,000 installations. While the dispute is open, a designer should set decay and aggregation so a trace serves coordination now without staying legible for later monitoring. The strongest objection is that symmetry is not the software's to set if the room does the coordinating.

Where the argument is heading

The area has moved from mechanism to relation. The first decade asked how to build awareness mechanisms, and produced a specification effort (Benford and Fahlen, 1993) and a framework (Gutwin and Greenberg, 2002). The decade after asked what the concept was, and produced an internal critique (Schmidt, 2002) and the argument that context cannot be sensed, only participated in (Dourish, 2001). Whether symmetry, decay and aggregation can be specified as system properties is the open question. Neither Ackerman and Starr's (1995) indicators nor Erickson and Kellogg's (2000) translucence has been tested against the monitoring outcomes the tradeoff predicts.

The sources this unit rests on

Each note states what the document did and found. Identifiers were resolved on 2 November 2026.

• Ackerman, M. S. and Starr, B. (1995). Social activity indicators: interface components for CSCW systems. doi:10.1145/215585.215969 Argues that indicators of social activity are usually ad hoc, and builds a generalised mechanism on a CSCW toolkit. It is aggregation in design form: not a log, an indication.
• Benford, S. and Fahlen, L. E. (1993). A spatial model of interaction in large virtual environments. doi:10.1007/978-94-011-2094-4_8 Presents a spatial model defining aura, nimbus, focus and adapters to control mutual levels of awareness. It makes awareness a quantity a designer sets, permitting the surveillance boundary to be designed.
• Cadiz, J. J., Venolia, G., Jancke, G. and Gupta, A. (2002). Designing and deploying an information awareness interface. doi:10.1145/587078.587122 Reports Sideshow, a peripheral awareness interface that 13,000 employees installed within a year and 7,000 used regularly; lost screen space and distraction were the main reasons given for stopping. It is rare evidence that awareness support reaches a population, and that users leave on their own terms.
• Carroll, J. M., Neale, D. C., Isenhour, P. L., Rosson, M. B. and McCrickard, D. S. (2003). Notification and awareness: synchronizing task-oriented collaborative activity. doi:10.1016/s1071-5819(03)00024-7 Analyses awareness breakdowns in a collaborative learning system and argues that notification supports awareness of presence and actions but not of persistent activities. Its activity awareness separates signalling an event from supporting an activity.
• Dourish, P. (2001). Seeking a foundation for context-aware computing. doi:10.1207/s15327051hci16234_07 Argues that context is a relational property produced through activity rather than a surrounding a system can sense. Hence this unit treats awareness information as interpreted rather than collected.
• Dourish, P. and Bellotti, V. M. E. (1992). Awareness and coordination in shared workspaces. doi:10.1145/143457.143468 Studies groups of three designers using the shared editor ShrEdit, distinguishing active awareness from awareness generated passively in the workspace. The designers complained they could see one another's input but not their cursors, so the passive trace carries the coordination.
• Erickson, T. D. and Kellogg, W. A. (2000). Social translucence: an approach to designing systems that support social processes. doi:10.1145/344949.345004 Argues that making participants and their activities visible to one another supports coherent behaviour, naming visibility, awareness and accountability as the traits of socially translucent systems, illustrated by the deployed system Babble. It makes visibility a matter of degree.
• Gutwin, C. A. and Greenberg, S. (2002). A descriptive framework of workspace awareness for real-time groupware. doi:10.1023/a:1021271517844 Develops a three-part framework for workspace awareness in real-time distributed groupware: the elements of knowledge, the perceptual mechanisms that maintain it and the ways people use it. It lets this unit treat awareness as specifiable rather than atmospheric.
• Heath, C. and Luff, P. (1992). Collaboration and control: crisis management and multimedia technology in London Underground Line Control Rooms. doi:10.1007/bf00752451 Examines the Line Control Rooms of London Underground, a complex multimedia environment in transition, explicating the tacit practices through which personnel communicate information and coordinate many tasks. Coordination is achieved through an arrangement, not a mechanism, which is this unit's third position.
• Hudson, S. E. and Smith, I. (1996). Techniques for addressing fundamental privacy and disruption tradeoffs in awareness support systems. doi:10.1145/240080.240295 Names a dual tradeoff between privacy and awareness and between awareness and disturbance, argues it is fundamental, and proposes four techniques that expose new design points by selecting what is transmitted. More awareness is therefore not a free improvement.
• Pinelle, D., Gutwin, C. A. and Greenberg, S. (2003). Task analysis for groupware usability evaluation. doi:10.1145/966930.966932 Develops Collaboration Usability Analysis, which models the teamwork in a group task rather than the taskwork, grounding collaborative actions in primitives called the mechanics of collaboration. It makes the coordination claim testable, since the evaluation must model the shared task.
• Prinz, W. (1999). NESSIE: an awareness environment for cooperative settings. doi:10.1007/978-94-011-4441-4_21 Describes NESSIE, application-independent awareness infrastructure with an open event protocol, configurable indicators and a server supporting access control, reciprocity and subscriptions by shared interest profiles. Reciprocity and access control become properties of the awareness layer, this unit's strongest support for symmetry.
• Schmidt, K. (2002). The problem with 'awareness': introductory remarks on 'awareness in CSCW'. doi:10.1023/a:1021272909573 Reconstructs how awareness entered CSCW as a placeholder for elusive practices of taking heed, records that the media space programme's expected benefits never materialised, and argues the term does not denote a set of related practices. Monitoring and displaying, it argues, are mutually dependent, which grounds this unit's first design variable.
• Tang, J. C. (1991). Findings from observational studies of collaborative work. doi:10.1016/0020-7373(91)90039-a Videotaped small groups at a shared drawing surface and found that hand gestures carried significant information, that drawing conveyed what the finished drawings did not, and that spatial orientation structured the activity. The artefact and the conduct around it are part of the coordination mechanism.
• Bly, S. A. (1988). A use of drawing surfaces in different collaborative settings. doi:10.1145/62266.62286 Studied two-person design sessions face to face, over an audio and video link, and by telephone, and found many of the same drawing activities in all three. It shows the act of drawing coordinates, and that a partial surface still carries much of it.
• Sum, C. M., Shi, C. and Fox, S. E. (2025). 'It's Always a Losing Game': how workers understand and resist surveillance technologies on the job. doi:10.1145/3710902 Analysed Reddit posts and worker interviews, finding layered surveillance met with thin transparency and everyday resistance.