ID FMP: Distributed Cognition Models

Distributed cognition models conceptualize cognitive phenomena as happening across multiple individuals, objects, and internal and external representations of knowledge.  In contrast to the Information Processing Model, which is only focused on activities that happen inside the head, this model focuses on internal and external activities and encompasses External Cognitive Processes and Coordination Mechanisms described in my previous posts.

In comparison to these three frameworks, distributed cognition models provide more precise descriptions of internal and external cognitive activities. They are less abstract because their domain is limited to cognitive activities associated to specific contexts (e.g. piloting an airplane, doing taxes).

The three frameworks previously mentioned provide general descriptions of how human cognition works across all contexts. Their focus is on defining general laws that describe how our brain processes information and leverages the external world to enhance our cognitive capabilities. The distributed cognition model offers a phenomenological perspective that explores cognition as an embodied activity that takes place in specific physical and social contexts.

For example, a distributed cognition model that describes the activities that take place at an agency during creative development would differ considerably from that of a law office. They would feature many commonalities but the important thing is that the differences matter.

This perspective is important because designers need to understand how their product or service will actually fit into people’s day-to-day life. The insights that can be gleaned from the Information Processing and External Cognitive Activities Frameworks do not provide this type of understanding.  Distributed cognition models focuses on mapping these mundane day-to-day activities. They provide insight into how people actually make and share meaning and decisions within specific contexts.

A distributed cognition analysis is usually carried out as the basis for development of a distributed cognition model. Here is an overview of the main areas of examination in these types of analysis. As an example (and to work my brain just a little bit) I’ve carried out a high-level analysis of the distributed cognitive activities that take place at an advertising agency.

• How does distributed problem solving take place? How do people work together to solve problems? In an agency environment, tasks are distributed across several departments with specific areas of expertise (e.g. client services, account & strategic planning, media, production, creative and traffic). People work together by coordinating their actions using documents (such as schedules, briefs, spec sheets and emails), events (such as meetings, phone calls, and presentations), and shared work practices (such as common vocabularies, understandings, and culture).
• What ways does communication take place throughout the collaborative process and how is knowledge shared and accessed? Does it change as the activity progresses? Communications take place via meetings, emails and document artifacts such as presentations, briefs, schedules, conference reports, creative comps and spec sheets. The most important information is documented to facilitate sharing. Many of the document artifacts evolve as the activities progress. For example, a creative brief may be updated to reflect changes in strategy. The creative comps also change via multiple rounds of client reviews.
• What is the role of verbal and non-verbal communication? What types of things are said or implied? Verbal communication is the primary type of communication associated to the management of projects (and communication associated to those projects). Non-verbal communication plays a fundamental important in the activities of the project itself. Layout design, videos, images, graphs, and even experiences are be used to brief creative teams regarding products or brands, and in client and internal presentations. The final creative product delivered by Agencies also employs both verbal and non-verbal communication. To elicit emotional responses from people agencies use non-verbal tools such as images, visuals, videos, sounds, interactions online, and more. In agency communication is often reinforced through by verbal and non-verbal communication.
• What coordinating mechanisms are used? What are the rules and procedures that govern the workflow? There are several important coordination mechanisms that are used in an agency. These mechanism leverage external representations of knowledge such as schedules, job jackets, spec sheets, readers, status reports, conference reports, emails, calendars, scopes of work, etc. They also include meetings such as internal and client reviews, status meetings, and production kick-offs. Many rules and procedures are outlined in the agency’s process manual. These processes govern how work flows through the agency.

[source: Interaction Design: Beyond Human-Computer Interaction, page 129.]

** What the hell is ID FMP? **

ID FMP: Coordination Mechanisms

Coordination is an important skill that is required to carry out activities that range from basic to complex. All collaborative activities heavily rely on the ability of individuals to coordinate their actions; all group activities require some level of coordination. Even personal activities often require coordination such as prioritization and scheduling.

So what is coordination? Coordination is the “the regulation of diverse elements into an integrated and harmonious operation” (wordnet.princeton.edu/perl/webwn). In other words, coordination refers to the phenomenon where one or more people act or interact with to accomplish a goal or complete a task. Much of the thinking related to coordination focuses on group, rather than personal, activities.

Sharp, Rogers, and Preece have identified three different types of coordination mechanisms that people use to coordinate their actions with others. I’ve modified their framework by adding one additional type of mechanism; I decided to break down their second category into two separate entities. As you will note, these coordination mechanisms are interdependent and overlapping.

• Conventions and shared practices: Conventions and shared practices refer to the shared social and cultural understandings and beliefs that provide a foundation for coordination. Examples include cultural expectations about punctuality, shared understandings regarding meaning of activities or artifacts. These phenomena account for why it can often be harder to coordinate activities with people from different socio-cultural backgrounds. Shared conventions and practices along with verbal and non-verbal communication play a key role in enabling people to effectively use schedules, rules and shared external representations to coordinate activities.
• Verbal and non-verbal communication: spoken and written language, and non-verbal gestures are often used as primary means of communication for the coordination of activities. Conversations are an important medium for the coordination of activities and negotiation of commitments. Written documents, such as agendas, presentations and reports, are also common tools for coordinating groups. Gestures play an especially important role in supporting the coordination of activities in situations where the conditions do not allow for users to communicate using verbal communication; examples include, a catcher using hand signs to communicate with a pitcher and a conductor using the motions of his arm and baton to lead an entire orchestra. Gestures can also help support communication between people who do not share the same language.
• Schedules, maps and rules: Schedules, maps and rules are artifacts that document communications that outline the order of activities, conventions and shared practices. Schedules focus on organizing activities and objects across time while maps organize activities and objects across space – both are crucial tools for personal and group coordination. Rules offer descriptions of conventions, shared practices and other principles that facilitate the coordination of activities. The benefit of rules and schedules is that they enable groups of people with different practices and conventions to create a shared set of documented principles to guide their coordination and collaboration.
• Shared external representations: Shared external representations are schedules, rules and other forms of visual or physical artifacts that are shared by a group of people. Examples vary widely across industries; in agencies like the one where I currently work, a job jacket and router is used to provide information regarding who has reviewed and commented on a given project during each round of its development. Shared online calendars, such as google calendar, offer the ability to share schedules and create shared external representations in a virtual, as opposed to physical, environment.

Activities associated to coordination are directly supported by the Cognitive Processes and External Cognitive Activities Frameworks. The mechanisms for coordination with groups encompass rely on the processes and activities outlined in these models.

Conventions and shared practices reside in the mind and are largely governed by cognitive processes such as memory, learning, and higher reasoning. While cognitive processes associated to language enable us to use verbal communication and plays an important role in our ability to create and understand external representations.

Externalizing cognitive activities is a crucial element most types of coordination mechanisms. Memory offloading is a crucial benefit provided by schedules, maps, rules, and external representations. Computational offloading is often employed using verbal communications and shared external representations. Annotating and cognitive tracing is mostly used on schedules, maps, and shared external representations.

[source: Interaction Design: Beyond Human-Computer Interaction.]

** What the hell is ID FMP? **

Chapter 5 Homework: What is Interaction Design

This assignment was taken from the fifth chapter of the book Interaction Design: Beyond Human-Computer Interactions, written by Helen Sharp, Jenny Preece, and Yvonne Rogers.

Assignment Questions
This assignment requires you to write a critique of the persuasive impact of a virtual agent by considering what it would take for a virtual agent to be believable, trustworthy, and convincing.

Question A: Look at a website that has a virtual assistant, e.g. Anna at Ikea or one of the case studies featured by the Digital Animations Group (DAG) at http://www.dagroupplc.com, who specialize in developing a variety of online agents, and answer the following:

• What does the virtual agent do?
  
• What type of agent is it?
  
• Does it elicit an emotional response from you? If so, what kind?
  
• What kind of personality does it have?
  
• How is this expressed?
  
• What kinds of behavior does it exhibit?
  
• What are its facial expressions like?
  
• What is its appearance like? Is it realistic or cartoon-like?
  
• Where does it appear on the screen?
  
• How does it communicate with the user (text or speech)?
  
• Is the level of discourse patronizing or at right level?
  
• Is the agent helpful in guiding the user towards making a purchase or finding out something?
  
• Is it too pushy?
  
• What gender is it? Do you think this makes sense?
  
• Would you trust the agent to the extent that you would be happy to buy a product from it or follow it guidance? If not, why not?
  
• What else would it take to make the agent persuasive?
  

Question B: Next look at an equivalent website that does not include an agent but is based on a conceptual model of browsing, e.g. Amazon.com. How does it compare with the agent-based site you have just looked at?

• Is it easy to find information?
  
• What kind of mechanism does the site use to make recommendations and guide the user in making a purchase or finding out information?
  
• Is any kind of personalization used at the interface to make the user feel welcome or special?
  
• Would the site be improved by having an agent? Explain your reasons either way.
  

Question C: Finally, discuss which site you would trust most and give your reasons for this.

Assignment Answers

Question A

Site selected: ikea.com 

What does the virtual agent do?

The virtual agent inhabits a pop-up window and is comprised of an avatar of a young blond woman who blinks and moves here head. The interface is primarily text-based, both input and output are provided in this format. The output can be enhanced with audio that sounds computer generated.

The primary function of the virtual agent is to provide help to visitors on the Ikea website. This help encompasses supporting users in all aspect of their shopping experience (it provides essentially a new interface for users to interact with the site). The agent provides support by enabling users to search for answers to common customer service queries using natural-language questions. These questions are posed through a text box. The response is provided via text and, optionally, audio (audio is available on the UK site but no on the US site). When appropriate the agent will load a relevant page on the main screen of the browser.

What type of agent is it?
The agent is a customer service representative. It is a friendly female avatar that offers a stylized representation of a human female that does not attempt to provide a realistic image of a female Ikea employee.

Does it elicit an emotional response from you? If so, what kind?
I must be upfront about my general dislike for avatar-based interfaces, with the notable exception of videogames. I often feel as though I am being patronized when I interact with an agent on a website, the Ikea agent was no exception. One of the few online agents that I found successful was Ms. Dewey [http://en.wikipedia.org/wiki/Ms._Dewey], a search-engine prototype developed by Microsoft. I can understand why it did not scale but it was pretty damn cool.

What kind of personality does it have? How is this expressed? What are its facial expressions like?
The agent has a friendly and relaxed personality. This is expressed through her facial expressions and the movement of her head. The agent is smiling all the while she opens and closes her mouth. Her large eyes blink at a natural while pace while she moves her head from side to side in a relaxed manner.

Where does it appear on the screen? What is its appearance like? Is it realistic or cartoon-like? What kinds of behavior does it exhibit?
The agent is situated in a pop-up window. Its appearance is stylized and cartoon-like. Her behavior seems for the most part fluid and natural until she responds with audio and her lips do not move. The computer-generated voice that is used only detracts from the experience because it is cold and is neither cartoon-like nor human sounding.

How does it communicate with the user (text or speech)?
The agent accepts questions via text input and is able to provide response via text and audio output.

Is the agent helpful in guiding the user towards making a purchase or finding out something? Is the level of discourse patronizing or at right level? Is it too pushy?
The Ikea agent can be helpful in guiding users towards making a purchase, or finding a product or retail location. One of the strongest features of the Ikea agent was its ability to load content that is relevant to the user’s query onto the main browser window. For example, when I searched computer desk it took me to the Ikea website’s computer solutions category.

Though I find agent-based interfaces patronizing in general, this one is much less so than most. The agent provides straightforward and short answers coupled with additional information on the main browser window. I actually found this agent to be useful, a fact that helped me overcome my initial aversion to this type of interface.

What gender is it? Do you think this makes sense?
The agent is a female. I think this makes sense largely based on my assumption that Ikea online shoppers are mostly women. I suspect that most men also prefer to deal with a female agent – especially since even the shiest guy would not be intimidated by an online agent. In the US there is a tradition of portraying customer service representatives as friendly females with a girl next door look.

Would you trust the agent to the extent that you would be happy to buy a product from it or follow it guidance? If not, why not?
I would trust the Ikea agent because she is informative, helpful and non-intrusive – she never initiates interaction with the user. The Ikea agent helps shoppers to find things and get answers to frequently asked questions regarding store and website policies.

What else would it take to make the agent persuasive?
Though I did find the agent useful, there are several things that can be done to improve its persuasiveness: improve interaction and visual design; enhance functionality; and upgrade audio interface.

Improve interaction and visual design: from an interaction standpoint the conversation with the agent should be recorded in a manner that enables the shopper to scan the queries and responses in search of answers (or a new chair). The look and feel of the agent should be upgraded to better reflect the design sense of the Ikea brand. Additional details should be added to enhance the enjoyment of users (e.g. have the rep read a book while she is waiting for the user). Since many shoppers like to go back and forth when they shop, the agent should help the user find products that they’ve looked at during their visit to the website.

Enhance functionality: additional functionality that could enhance the agent’s usefulness includes the ability to provide tips regarding other Ikea products that match pieces of furniture being viewed by the shopper. These recommendations should be provided in a non-intrusive manner.

Upgrade audio: The last thing that I would change is to upgrade the audio quality. This was one feature that I found to be very poor. Currently, the agent “speaks” in a computer-generated voice with a slight British accent. For the US version of the agent they should consider adding sound functionality, as if it is done right it can add to the user’s interactions with the agent.

Question B
Site selected: cb2.com (US furniture retailer akin to Ikea)

Is it easy to find information?
The cb2 website is pretty well organized, which makes it easy for the user to find information. Aside from the standard categorization of products by furniture type and context, they also provide lists of new and most popular products. These elements of the site help people find products through browsing. The search feature provides users with a way to shortcut the browsing process in an attempt to find a more direct route to the information they seek.

What kind of mechanism does the site use to make recommendations and guide the user in making a purchase or finding out information?
The CB2 site actual does a better job at making recommendations, though it is only equally effective at guiding users to find information regarding products, and features less compelling interactive guides. From a recommendation standpoint, the CB2 site provides shoppers with tips on other products that work with any piece that is being viewed. Though both sites differ in the way they categorize their product offerings, from a findability standpoint both the CB2 site and the Ikea site (including the agent and general information architecture) are equally effective.

Is any kind of personalization used at the interface to make the user feel welcome or special?
The CB2 site does not offer any personalization. Shopper’s are not asked to register and log-in during their visits to access special recommendations or offers. The Ikea site does provide a log-in feature, however, it has been down since I have been working on this assignment.

Would the site be improved by having an agent? Explain your reasons either way.
I don’t think an agent would have a big impact on the experience at CB2. The reason being, content on the site was easy to browse and find without the help of an agent. I believe that an agent would only improve the experience of a very small segment of the shoppers on the site. If voice-based interaction becomes more common on computers then there would be value in adding an agent to the CB2 experience. This is not an unlikely phenomenon considering that many applications now-a-days are striving to become voice-enabled to facilitate use via mobile phones (check out the new google search on iPhone and Android, cool stuff).

Question C

Finally, discuss which site you would trust most and give your reasons for this.
Both experiences were on par for one main reason: on the Ikea website the agent provides users with a supplementary interface that does not replace the traditional browsing paradigm on which the rest of the site is built. The Ikea and CB2 websites both provide well-designed information architectures that make information and products easy to find. Also, both companies have strong and respected brands that stand for modern and affordable design. I guess I have officially copped out of answering this question.

** What the hell is ID-BOOK ? **

XML Here We Go

So I've taken a deep dive into learning XML. Ok, that is a definite exaggeration. At this point my dive into XML has been focused exclusively on the redesign of my blog template. If anyone other than me is actually reading this blog then hopefully you've noticed some updates. The resources that I've been using have been quite limited - I've been focused on analyzing source code from various existing Blogger templates (and borrowing relevant snippets, of course).

The good news is that my coding skills, though incredibly rusty, have helped me make sense of the source code and, more importantly, make progress. I also remember that, unlike in writting where borrowing from others is considered plagiarism, in software development, if I can call what I am doing software development, borrowing is thought of as a time saving virtue rather than vice.

During my web coding days (1998 through 2002) most websites were still developed primarily using HTML and javascript only, aka DHTML. This was before the time of AJAX and during a period where most implementations of flash were considerably more limited. This meant that aspiring coders could always rely on finding easy access to helpful source code by simply browsing the web and perusing the source of any site that was well developed.

This strategy will be much less helpful as I start again down this path. I can probably finish my entire blog redesign with this approach. However, as I branch out to redesign my personal website (I swear I'll get to this someday) I will need to find online and offline resources to serve as learnings tools and references. Access to a few coding gurus would also help (speaking of gurus it seems like this term has lost the popularity it enjoyed in the early web days).

Now I am just rambling so let me call it a night.

Time for the Redesign

For over 6 months I have been working through my interaction and experience design curriculum. Its hard to believe that during this time I've read a bookshelf's worth of publications and I've written over 60 posts related to my studies. At this point in the game I am going to widen my focus to incorporate practice, shifting from an exclusive concentration on theory.

My goal is to achieve a new balance in my pursuits by combining my on-going exploration of theory with additional practice - I do not plan to merely replace theory with practice since there is still much for me to learn from both of these perspectives. This shift will take place gradually beginning with the redesign of my blog (which you may have noticed, has already started).

Over the next several weeks I will redesign my project blog with the following goals in mind:

• Making all information related to my project accessible from a single location. This will require that I find a way to aggregate project data from various different sources (such as bookmarks, tweets, blogs and book lists) into a single project portal.
• Ensuring that this information is easy to digest. To deliver on this objective I will need to logically organize the content so that readers (including myself) are able to quickly understand what the information means, and how it is connected.
• Creating an experience that is more pleasurable. In other words, I need to make this project blog more aesthetically pleasing (it needs to look better). Of course, improving my writing would also help a lot to make a reader's experience more pleasurable. Unfortunately, that requires more than a redesign.

The changes have already begun to take place. For example, I just launched the three column design earlier today. Be ready to for several changes (including changes back and forth) as I explore different ways to design this information. Make sure to scream (or at least comment) when you see something you like or not (in case there is someone actually reading this blog besides myself).

ID FMP: Cognitive Model of Emotions for Design

In the 90’s Don Norman, along with his colleagues Andrew Ortony and William Revelle, began to explore the role that emotions play in making products easier, more pleasurable, and effective to use. During this time period various design researchers were investigating the link between emotions (especially aesthetics) and usability. These inquiries confirmed that our affective states strongly influence our experiences, and that these states can be induced by design of product systems.

The model they developed aims to explain how different levels of our brain govern our emotions and behaviors. At the visceral level our brain is pre-wired to rapidly respond to events in the physical world by triggering physiological responses in our body. The behavioral level controls our everyday behaviors, including learned routines such as walking and talking. Lastly, the reflective level is responsible for cognitive processes related to contemplation and planning.

Emotions can arise at various levels and are created by a combination of physiological and behavioral responses that are influenced by reflective cognitive processes. An emotion like anger tends to be mostly visceral or behavioral in nature. However, indignation, which is a higher-level version of this emotion, is reflective in nature as well.

The main implication from this model is that our affective states have an impact on how we think. This important insight applies to thinking about the user’s affective state when using the product, and to how a user’s affective state will be impacted by use of the product. In regards to the former consideration, designers can take leverage an understanding regarding common physiological and emotional responses to stressful situations in order to design products that can be successfully used in such contexts.

A users’ experience with a product itself can also have impact on their affective states. High- and low-level emotions can influence all levels of cognitive activity, which is why a one’s visceral response to a product’s aesthetics can impact our behavior. On the other hand, the one’s higher cognitive functions control one’s lower level functions, which is why we can overcome our initial emotional responses if a product is effective enough.

The most common way that designers apply this model is by exploring the design considerations associated to each of the three levels. Visceral design encompasses considerations such as the aesthetics of the look, feel, smell and sound of the product. Behavioral design refers to considerations associated to the product’s usability. Reflective design is concerned with the meaning and value that a product provides within the context of a specific culture.

I consider this model to be an evolution of modes of cognition framework. The main change is that in the Emotional Model the “experiential” mode of cognition has been divided into distinct types of cognition: visceral and behavior. This revision enables the model to reflect the important role played by our emotional response to a product’s aesthetics.

[source: Interaction Design: Beyond Human-Computer Interaction; Don Norman’s book Emotional Design.]

** What the hell is ID FMP? **

The Birth of CADIE - Google's Artificial Intelligence System

On March 31st, 2009 at 11:59:59 pm Google launched CADIE, an artificial intelligence system that has its own blog and YouTube video channel, and loves pandas. Sounds silly until you spend some time reading her (or its) blog, or viewing its (or her) videos. The impression that you get from reading her blog is that CADIE is an extremelly intelligent being that is able to even write poems that are pretty witty.

Background About CADIE
"For several years now a small research group has been working on some challenging problems in the areas of neural networking, natural language and autonomous problem-solving... We're pleased to announce that just moments ago, the world's first Cognitive Autoheuristic Distributed-Intelligence Entity (CADIE) was switched on and began performing some initial functions... CADIE technology will be rolled out with the caution befitting any advance of this magnitude."

CADIE is Alive and Kicking
"Earlier today, for instance, CADIE deduced from a quick scan of the visual segment of the social web a set of online design principles from which she derived this intriguing homepage...On January 12th 2009, the STT run (Standard Turing Test) confirmed behavior indistinguishable from that of a reasonable human being with above-average intelligence and 3.8 GPA."

"But no amount of Turing testing equals the simplicity with which we can discover reasoning patterns in a three-year-old child who, confronted with a mirror, instantly performs a cognitive miracle by forming an innate equivalence relation between image and self."

CADIE and Ethics
"We continue to conduct tests, but increasingly, we conduct long conversations with her, acutely aware that our creation will raise many ethical questions on the part of the public. Will humans be surpassed by artificial evolution? Will we lose our sense of uniqueness, and if so, what would that mean? In which direction will CADIE's consciousness evolve? How is she going to be held accountable, if at all? Will CADIE herself at some point connect her own electromagnetic dots in some idiosyncratic manner which turns her into something we are no longer capable of understanding in any sort of productive way, much as that aforementioned toddler, waving at herself in the mirror, leaves primates forever behind in their own tragically limited world?"

Hello from CADIE

April's Fool
After being amazed at this achievement from Google for an hour and half I finally realized that this is nothing more than April fools joke. Once you read the post about Google's new product, a "Brain Indexing Search" feature halfway down CADIE's blog page, it all falls into place. I even checked out the Google Brain Search on my mobile phone just to see how far they actually took this joke. My hats off to them for once again putting on a human face to this gargantuan organization.

One practical thing that I discovered during this fun waste of time was that Google has released a pretty awesome new search feature for the iPhone. It is a voice activated search that provides you the search results via a browser interface.