IS 240 -- Information Systems Analysis and Design

Assignment for week 9: Information design

Information design concerns the visual display of structured information. It derives not from the technology world but from graphic design: people who design the visual appearance and organization of things like brochures, signage, museum displays, and expository illustrations for newspapers and textbooks. We are emphasizing information design in this course because we believe that it is the future. In particular, the ergonomically oriented field of user interface design is stuck in the single windows/menus/mouse paradigm, and the more general field of interaction design does not have adequate ideas about the organization of the information that one interacts with. We do believe that these fields have their place, but that the best first step is to learn from the graphic designers about the visual display of information before revisiting the question of how to interact with it. So in asking you to conceptualize the information design of your service, we are not asking you to figure out what commands go in what menus. Rather, we are asking you to figure out whether you want to use menus, or whether you want to draw on some other information design metaphor. More generally, our focus is not on interaction but on the organizing principles of the things one interacts with.

Last week you worked out some of the internal structure of your information service. Your assignment this week is to prototype the physical form and information design of your service. The purpose of this document is not to explain what information design -- the readings and lectures will do that. Rather, this document will present a conceptual framework for understanding the full scope of the problem that information design solves. Along the way we will also discuss issues of physical form, although that is a secondary focus.

The first priority of information design for electronic information services is that the whole design, inside and outside, should be unified by a single conception: a clear and powerful metaphor or symbol or ideology that provides guidance to every aspect of the design. The most famous example of such a unified conception is the idea of individual liberation that informed the design of the Apple Macintosh. The Macintosh was design in depth: not only was every aspect of the hardware, operating system, packaging, and company communications informed by the liberation ideology, but the designers also wrote a manual for the "look and feel" of Macintosh applications that were developed by third parties. Another, closely related design principle of the Macintosh was a clear mapping between the internal workings of the machine and its outward behavior. Entities of the data model became types of record structures, and they also became elements of the interface. Visual elements (e.g., icons) and commands (e.g., menu items, behavior of the mouse) mapped cleanly onto the internal workings of the software. The increased power of computer processors and the increased diversity of input and output devices and communications mechanisms have greatly multiplied the design space for interactions between people and machines, but these design principles still hold. Of course, it may be impossible to achieve the idea of a unified conception that drives every aspect of a design. But it is important to try, and every aspect of a design that diverges from the unifying conception is likely to be problematic in some way.

As the design space widens, it helps to have a conceptual framework with which to map the full range of potential information designs. Among the dimensions to consider are these:

* The framed image, independent of context and interaction. Does it work aesthetically? Does it make sense? Is it illegible? Cluttered? Confusing? How does it work as a diagram? A sign? An advertisement? A work of art?

* Time. How does the image change over time? Does it change in response to actions that people take, and if so which ones? We are accustomed to the personal computer model and the "user illusion", according to which almost all changes take place in response to specific commands by the person who is sitting at the keyboard. But in a distributed system that interconnects many people who occupy several differentiated roles, the design space is obviously much larger.

* Narrative. If the information participates in telling a story, explaining something, simulating something, or showing how something works, then all of the categories of narrative apply. A narrative has a beginning and end, a world with rules, an internal logic, characters and places, events and scenes, and so on. It also has instructed seeing: not just the "showing" aspect of the story (facts, events, actions, and other concrete specifics) but the "telling" (what it meant, what was really happening, the pattern, what the character was thinking, what aspect of the events are to be significant, and so on). Even when the information service or the people who are using it are not clearly telling a story, narrative concepts can be heuristically helpful in articulating the temporal dimension of its relationship to the people who use it.

* Action. Systems analysis produces, among other things, a repertoire of actions that people in various roles need to be able to take by using the service. Examples of actions might be bidding on a product, signing off on an invoice, deleting a message, reclassifying a document, promoting someone to a new job title, or closing a case. Notice that some of the actions are defined purely in terms of the service's workings (e.g., deleting a message), while others (bidding, signing off, reclassifying, promoting, closing) are also institutionally significant -- that is, they change the world and not just the machine. The visual display of the service might make clear which actions are possible at a given juncture. It might provide feedback about which action has been taken. It might ask for reconfirmation. It might suggest actions. Although our focus here is primarily on the visual display of actions (e.g., with icons that can be clicked), one might also accomplish these purposes using sound (e.g., recognizing or speaking words) or touch (e.g., the "clicking" sensation of a key), and you are welcome to mix the various modalities if that's the direction the design process takes you. Observe that the visual representation of action is usually closely related to the visual display of everything else; this is particularly evident in the case of "direct manipulation" interfaces such as the dragging and dropping of icons in the Macintosh.

* Linguistics. Human language is the starting point of systems analysis, and in most information design human language is selected and associated with other meaning elements. (Even when no words appear on display screens, you should be aware of the language that people will use in thinking and talking about your service.) This language should be consistent in both the words you choose and the meanings you assign them; it should be appropriate to the people who are using the service; it should avoid trampling on other meanings that the same words might have in contexts that the people already know about; and it should be aligned in every possible way with the unifying conception of the design. It should be easy to configure the service to use a different language or dialect. Linguistic elements can be wildly misinterpreted, and they should always be tested with real members of the relevant communities -- especially ones who are less familiar with services such as yours -- to see what misinterpretations your chosen words might inadvertently invite.

* Design language. A "design language" is not a set of words of natural language. Rather, it is a set of design elements and design conventions that are used consistently in a range of contexts. It is crucial to have a design language any time that more than one visual display of information is supposed to operate as a "family". For example, all Macintosh applications are visibly and unmistakeably *Macintosh* applications. Every page generated by a given Web site should likewise employ shared conventions. The signs accompanying a museum display should be come in visually distinct categories (text explaining whole a room as opposed to an individual picture) while also clearly being part of the overall information design of the show. The elements in a design language might include symbols (e.g., icons, logos, navigational devices), typefaces (e.g., Times Roman for everything, italics for captions, small type for legal boilerplate), colors (e.g., background, pull quotes, highlights, selected items, urgency), and vocabulary items. The design conventions might include the layout of information on the screen (e.g., commands on the left, site map at the top, conventions from paper genres like spreadsheets), the set of actions that are available and how one performs them (e.g., drag and drop, "OK" button in the lower right corner), the use of titles and summaries (e.g., automatically generated text abstracts set off in side-bars), and the spatial relationships among the design elements (e.g., text is always on the button rather than next to it).

* Structure. Most information designs will include both structured and unstructured elements of a document, and each poses its own design challenge. For structured information, the challenge is to make the structure visible, while also making visible the lesson or point that can be taken home from the structured information as a whole. A tabular display of data, for example, does not make trends as visible as a graph does. But at least structured information occupies a predictable amount of real estate on a visual display. Unstructured information such as free text can be less predictable, and you may need a style manual for writers that dictates the size of each textual element such as a caption. Images are predictable in their spatial demands, but it can be very difficult to direct someone's attention within an image. Actions over structured information can be more complex than those defined over unstructured information.

* Reference. Every visual display is part of history, and it cannot help referring to culture, tradition, genre, and style. Everybody who uses your service will already have mastered a staggering number and variety of visual codes, whether from technical diagrams, films, graphic design, cartoons, or computer interfaces, and the visual displays that your service generates will inevitably be interpreted with reference to those codes. If your service is going to remind your community of a textbook, then you must either embrace that reference or revise your design to stop inviting it. This principle applies on every level: from individual design elements, to conventions of spatial layout, to ways of telling a story, to ways of taking an action, to literary genres. If your service resembles the Macintosh interface then you will have little choice but to uphold the conventions of that interface. If your service looks like a comic book or a Russian propaganda poster or a television screen or a Rand McNally road map, then you have to anticipate that interpretation and take a stand toward it.

* Embedding. Your visual display does not exist in a vacuum. It is a physical object that stands in some physical and practical relationship to everything else that is going on around it. Think, for example, of a display that is a sign that people walk past, or a display that is sensitive to the devices that might be present, or to environmental conditions. People might look at the display from a variety of distances or angles, or they may notice changes in the display from the corner of their eye. Lighting conditions may not be reliable. Glare may be a problem.

* Attention. Closely related is the structure of the individuals' attention. What else is going on when people are looking at the display? Is the display calling out for attention or competing for attention with other visual things? Is the individual focused on the display for the whole time that the service is telling some story, or does the service have no guarantee about the times that people might arrive or leave? Is the individual's relationship to the display mainly exploratory or routine or what? Are multiple individuals engaged in joint attention, for example with one person explaining something to another person by directing their attention within the display? Will the individual's attention be directed by changes in the display that are caused by someone else at a remote location, for example in a distributed game or dialogue? Will particular design elements within the display be inviting attention? And so on.

* Function. What function does display play in the ongoing activity? The ongoing activity will probably be structured by an institution, a set of customs, a relationship among people, some rules, and so on, and the display will take on most of its significance within this context. A menu, for example, serves a very stable function in the rituals of the restaurant.

We are also asking you to design and mock up the physical form of your design. This could include a keyboard on a destop, a flat screen display on a wall, a device that resembles a wristwatch or cell phone or palm pilot or credit card, something that is sewn into clothing or mounted on eyeglasses, installed in a car, carried on a belt or in a purse or backpack, or whatever you can imagine being built with the technology of ten years from now. You are most welcome to presuppose the processing capacity and communications capabilities of 2010, and to assume the kind of spontaneous wireless interaction among devices that we discussed in week 5. As with the information design of your service, the physical form should also express the overall conception of the design as a whole. It can do this in an infinite variety of ways, including its shape, markings, colors, and plugs, or in the way that a person holds it, interacts with it, notices it, shares it with others, hides it, discards it, and so on. Its cost may be significant. And just as the information design inevitably refers to a vast history, so the physical form will inevitably refer to a vast history of industrial design. Furthermore, "information design" and "physical form" overlap, inasmuch as the object itself conveys information whether it changes outwardly or not. Information design applies to simple LED's just as much as it applies to wall-sized bit-mapped displays, and it also applies to symbols and icons on the physical device, such as the ones that label the connectors, or that sit on the edge of the screen to show where the touch sensitive areas are. All need to be designed in a coherent way.

We are concerned particularly with information design that supports shared work, and so it may help to reflect on a number of examples of such design:

* "Local design" -- the design conventions evolved within a group, such as notation and vocabulary, whether wholly original or inflections or variations or additions upon conventional designs.

* Being aware of one another, like those video space interfaces that include "postage stamp" video of each other's office so you can see who is in.

* Planning and coordinating collaborative work, such as PERT charts -- in a networked world, those same displays can probably be continually updated to provide a running display of project status.

* Information about individuals, such as contact information, skill set, resume, professional network, whereabouts, current actions, history of actions, etc.

* Materials that individuals in the group prepare for one another, for example materials to review before a meeting.

* Hand gestures and other symbols that synchronously accompany interaction.

* Any representations the group members may be working together jointly to produce, and a wide range of annotations and suggestions etc about drafts of those deliverables.

* Agendas and minutes of meetings, and representations of parliamentary-type procedures like voting, deciding who speaks next, moving along to the next topic, etc.

* All sorts of speech acts -- even though these might be done with unstructured English in a face-to-face context, when done remotely or asynchronously they might require structure or design etc.

* Instructions (e.g., judge to jury, manager to subordinates, manufacturer to repair person) that a team might jointly use to organize its work together.

* To-do lists, lists of action items, and documentation of the work.

* Progress reports, evaluation and feedback, queries to others outside the group, etc.

* Representations of group work processes that are meant as advertising for the group and its work, or reports on the work in a journal.

There are probably many more categories of information design to support joint work, but that's a sampling anyway.

That brings us, finally, to the process you'll follow in your own information design work, and to the results you will show to the class. Like every other sort of design, information design is iterative. Do not get overly detailed in your first several iterations. It is much more important to have a clear concept, and to explore the various dimensions of its consequences, than to spell out all of the details, especially at the beginning. Just develop enough of the consequences of your concept to understand the problems and opportunities that arise from it, and then reconceptualize. Rather than dwelling on one design concept, develop a number of them and explore them all, picking the one that seems to be going somewhere. Remember, once again, that your information design concept needs to be aligned with the unifying concept of the whole design. You may find that the unifying concept that you developed last week does not suggest a compelling information design, and that a new information design concept suggests a reconceptualization of the whole function and structure of the service. This is a good thing, and you should embrace it.

In this assignment we are not emphasizing "user" participation. What you are developing is an early mock-up that illustrates a concept, not a serious candidate for a implementation as a finished system. You might think of your presentation to the class as a presentation to a community of practice that might use your service. Experience shows that such presentations work best when they are relatively low-tech. Slick PowerPoint demonstrations or carefully worked-out interfaces communicate your intention to implement just exactly what you designed. By contrast, mockups with cardboard and colored pens are nonthreatening, and because they connote a low level of effort they communicate a willingness to iterate, or even to throw things out and start over if necessary.