Winter 2013 - INF STD272-1 - LIEVROUW
Human/Computer Interaction

Week 3 | January 22 
Evolving Perspectives 2: Phenomenology, Ethnography, Speech Acts


Winograd, T. and Flores, F. (1986). Using computers: A direction for design. In Understanding Computers and Cognition: A New Foundation for Design, pp. 163-79. Norwood, NJ: Ablex. File


directions for the future
most important designing is ontological
ontologically oriented design is both reflexive and political

relevance of theoretical orientation to design of computer systems
•
o Using coordination systems as primary example
o Intended scope is larger: all technology
• Phenomenological insights about
o Readiness-to-hand
• Transparency of interaction is important in design of tools, but not best achieved by mimicking human faculties
 Car
• You’re driving down the road, not operating controls, phenomenologically
• Long evolution of design
• Providing the right coupling between driver and action in the relevant domain
 Computer interface
• Special function keys
• Pointing devices
• Menus
• Domains in which actions are generated and interpreted – more important
 Bad design forces user to deal w complexities that are in the wrong domain
 Successful system builders learn to consider user domain of understanding
 More systematic principle: programmer designs the language that creates the world in which user operates
• Can be ontologically clean or a jumble of related domains
• Clearly and consciously organized ontology is the basis for the kind of simplicity that makes systems usable
o User is driving, not commanding
o MT: what about the fact that cars now want you to talk to them?
Next step: move beyond superficial structures of words and pictures into the domains generated by what people are doing when they manipulate those structures
o Breakdown
• Central role of breakdown in human understanding (Heidegger)
 A situation of non-obviousness
• Recognition that sthg is missing leads to unconcealing some aspect of the network of tools
 Reveals the nexus of relations necessary to accomplish task
• Creates clear design objective: anticipate breakdowns and provide a space of possibilities for action
o Aids
• Training
• Understanding of domain
• Development of skills and procedures for coping
• Plays a fundamental role in design
 Domain of action
• Objects and properties that emerge in breakdown
 Basis of heuristic methodology often followed by good programmers
• Must anticipate the range of occurrences outside normal functioning and provide means to understand and to act
• Analysis of a human context of activity can begin with domains of breakdown, which can then be used to generate the objects, properties, actions that make up the domain
o Blindness
• In providing a tool, we will change the nature of how people act
• Often, no clear problem to be solved; situation of irresolution
o Opens opportunities for action
o The space of possibilities is determined by the particular offerings and the features they exhibit
• A business (org) is constituted as a network of recurrent conversations
o Each has its own structure of completion and states of incompletion with associated time constraints
• Conversations are linked in regular patterns of triggering and breakdown
o Goal of conversation analysis is an explicit description of the linkages between recurrent conversations
• Normal triggering
• Breakdown triggering
o Description makes it possible to see possibilities for restructuring the network on the basis of where conversations fail to be completed satisfactorily
• Creating tools mean designing new conversations and connections
o Most significant innovation in change is the modification of the conversation structure
o Changes of devices and conversation structure go hand in hand, so changes in structure may not be noticed
• Design includes the generation of new possibilities
o Careful analysis of conversation structure can help reveal conversations with potential for expansion
o Computer’s power comes from its connection to the larger network of communication in which organizations operate
• Domains are generated by the space of potential breakdown of action
o Much computer automation deals with standard domains (accounting, billing, scheduling, etc.)
o Ultimately the present-at-hand world of objects always is based on the breakdown of action
o MT: what is you just want to be more efficient. No breakdown, just feel like working fewer hours or sthg
o Grounding of description in action pervades all attempts to formalize the world into a linguistic structure of objects, properties, and events
o System development methodologies have a fundamental condition for generating the relevant domains: gradual emergence in practice
• Breakdown is an interpretation—everything exists as interpretation within a background
o The background shared by a community is what makes individual tastes/preferences possible
o Satisfaction is determined by a declaration on the part of the requestor that a condition is satisfied
• By an individual, within the background of a community
• Domains of anticipation are incomplete
o Every attempt to anticipate breakdown reflects a particular domain of anticipation
• Must design with flexibility to encounter unanticipated breakdowns
• Conversations that depend on the assumption that a declaration of an interpretation is a fact will fail to deal with unexpected cases
• Computers are a tool for conducting the network of conversations
o Embodying possibilities for action within recurrent conversations
o Computerization: devices that were designed without appropriate consideration of the conversational structures they engender
• Innovations have their own domains of breakdown
o Breakdowns in the larger “web of computing” in which the equipment resides
• Design always is happening already
o Orientation that leads to asking significant questions, not a magic solution
o Design always proceeds, with or without an articulated theory
Systematic domains
• Language doesn’t describe the pre-existing world, but creates the world about which it speaks
• Distinction is different from explicit formal articulation
o But whenever there is a recurrent pattern of breakdown, we can choose to explicitly specify a systematic domain for which definitions and rules are articulated
o Computers are wonderful devices for the rule-governed manipulation of formal representations
• In applying computers appropriately to systematic domains, we develop effective tools
o Systematic domains
• Ain’t new (math)
• Profession-oriented domains
 Creation of expert systems are the creation of systematic domains that are relevant and useful to a particular profession
 Can be the basis for
• Computational tools that do some tasks previously done by professionals
• Tools that aid in communication and the cooperative accumulation of knowledge
• A structured formal representation that deals with things the professional already knows how to work with, providing for precise and unambiguous description and manipulation
 The critical issue is its correspondence to a domain that is ready-to-hand for those who will use it
• Common elements of profession-oriented domains
 The role of language in coordinated action
• Computer is a structured dynamic communication medium
• Communication is not a process of transmitting information or symbols, but one of commitment and interpretation
• Specification of mechanisms
 Kind of objects, properties, relations are suited to describing them and that can be embodied in a systematic domain
 Programming languages in general are not suited to the communication of intent and conceptual structure
• Too oriented to the structure of the machine, not the structure of its behavior
• System specification languages emerging that deal with the domain of computational devices in a more general way
• Central task is to characterize the precise form and relevance of the domain within a broader orientation
 Interaction between contextual and the systematic
• Leaving the hard questions open to human interpretation
Technology and transformation
• Symmetry: the world determines what we can do and what we do determines our world
• Activity evolves in accord with the structure of the new domain
o Hermeneutic orientation and biological theories provide insight into this process
• The key to cognition is the plasticity of the cognitive system, giving it the power of structural coupling
o As domain of interactions is modified, the structure of the interacting system changes
• Whenever we design changes to the space of interactions, we trigger changes in individual structure
 Changes to the horizon that is the precondition for understanding
• Computers have powerful impact because they are machines for acting in language
o Pre-understanding reflects the rationalistic tradition
• Authors criticize
 Biases about objectivity, nature of facts and their origin, role of the individual interacting with the computer
 As we work with devices with domains of action based on an interpretation of data, goals, operators, etc., we develop patterns of language and action that reflect these assumptions
• These patterns carry over into our understanding of ourselves and our behavior
 Descriptions of human thought as decision making and language understanding as the manipulation of representations reflects a deep concern with the discourse and actions generated by a rationalistic interpretation of human action
• Computer systems can reinforce this interpretation and working with them can reinforce patterns of acting that are consistent with it
• Opportunties
o Create computer systems that lead to better domains of interpretation
• Coaching in which new possibilities for interpretation and action emerge
• MT: this kind of coaching can apply to curiosity, as well
• Commitment is the basis for language
• Computers (technology) are a vehicle the transformation of tradition
 Our actions are the perturbations that trigger the changes, but the nature of those changes is not open to our prediction or control
 Nor can we objectively observe the transformation
 Continuing work toward revealing it is at the same time a source of concealment
In ontological designing we are doing more than asking what can be built. We’re engaging in a philosophical discourse about the self—about what we can do and what can be. Tools are fundamental to action, and through our actions we generate the world. The transformation we are concerned with is not a technical one, but a continuing evolution of how we understand our surroundings and ourselves—of how we continue becoming the beings that we are.

Suchman, L.A. (1990). What is human-machine interaction? In S.P. Robertson, W. Zachary, and J.B. Black (Eds.), Cognition, Computing, and Cooperation, pp. 25-55. Norwood, NJ: Ablex. File


What is human-machine interaction?
Interactive artifacts
• Notion
• Use
• Propagation
• Basis
Premise: interaction or communication turns on the extent to which my words and actions and yours are mutually intelligible
• Basis for beginning to speak of interaction, or mutual intelligibility, between humans and machines
Notion of h-m interaction
• Derives from the increased reactivity of computer-based artifacts
• The means for controlling computing machines and the behavior that results are increasingly linguistic rather than mechanistic
o Because of this shift, designers and users employ terms borrowed from the description of human communication
• Specific technical meanings within design, but maintain broad connotations of their ordinary usage
Interaction and instruction
• Designing and using artifacts is a problem of communication
o The idea of a self-explicating artifact
• Interactive machines are a solution that provides a tool with instruction in its use
• Face-to-face human interaction is designed for maximum context sensitivity and supports a response designed just for this recipient, just on this occasion
o Brings context sensitivity to bear on problems of skill acquisition
o Means that the response skills must be deployed anew each time
• MT: but isn’t easier as patterns emerge?
• Instruction manual is durable, reusable, replicable and allows for the disassociation of the occasion of an instruction’s production from the occasion of its use
 But affords relatively poor resources for recipient design
• Promise of interactive systems
o Technology that can move instructional design away from the written manual in the direction of the human coach and of the resources afforded by face-to-face interaction
• Self-explicating machines use the metaphor of machine as expert, user as novice
o Action interpretation based on a view of purposeful action that is central to work on machine intelligence and is deeply rooted in our common sense
• MT: how can it be universal, then, across cultures, education, religion, etc.?
• The meaning of our actions is best understood as the reflection of our underlying plans
 Project the user’s actions as the enactment of a plan, then use that plan as a template for the actual action’s interpretation
• MT: what if someone is just fucking around, or doesn’t know how to cause the action he/she desires? And is this template for all users? Or just that one?
• The machine’s function and design should serve as adequate context for the interpretation of the user’s actions, just as user’s purposes are constrained by the machine’s function
• Premise of the design is that the plan corresponds to the user’s actions and that the correspondence enables the interaction
o BUT, in actuality user and system each have a fundamentally different relationship to the plan
• The difference in the relation of user and system to the design plan has consequences for their interaction
• The coherence of the users’ actions (in this study) were largely unavailable to the system
• MT: a machine doesn’t understand! (p. 30)
• The strategy of binding certain user effects to particular machine responses enables the appearance of instructions occasioned by the user’s actions
o MT: assumptions here!!
o Machine only has access to users’ detectable actions
 Even though the system’s behavior changes, it can appear to behave in the same way to the user
• Accommodation to circumstance is fundamental in HCI
• Often a check of the current state of the machine belies the users’ intent
o MT: especially now, when people have greater access to machines without having specialized knowledge, training, experience, understanding
• Communicative troubles
o The design succeeds just in those cases where a detectable state can be linked to a prior assumption about the user’s intent with respect to a next action
• A single user action does not carry sufficient information to allow the attribution of its intent
• A consequence of the uncertain relation of expressed intent to situated action is an unavoidable uncertainty to shared understanding
 MT: shared understanding? How can that ever be? I think this is just an example of faulty modeling of human behavior in the programming!
 One approach: compile a catalog of frequently occurring misunderstandings and attempt to forestall them

Orlikowski, W. (1993). Learning from Notes: Organizational issues in groupware implementation. The Information Society, 9(3), 237-250. File

Dourish, P. (2004). What we talk about when we talk about context. Personal and Ubiquitous Computing, 8, 19-30. File

Boehner, K., DePaula, R., Dourish, P., and Sengers, P. (2005). Affect: From information to interaction. CC ’05: Proceedings of the 4th Decennial Conference on Critical Computing: Between Sense and Sensibility (Åarhus, Denmark), pp. 59-68. New York: ACM.