Activity Theory in Action
From CS260SP09
Readings
Activity-Based Prototyping of Ubicomp Applications for Long-Lived Everyday Human Activities Yang Li and James A. Landay, Proc. ACM CHI 2008.
Activity-based computing: support for mobility and collaboration in ubiquitous computing Jacob Bardram, Personal and Ubiquitous Computing 9, 2005.
CAAD: An Automatic Task Support System Tye Rattenbury and John Canny, Proc. ACM CHI 2008.
Discussions
Please post your critiques/commments on the required readings below. To do that, first login by using your user name and password, then click the "edit" tab on the top part of this page (between the "discussion" page and the "history" page), New to wikis? Read the Wiki editing guide. . Hint - Please put a whole line == ~~~~ == (literally) at the beginning of your submitted critique, so the wiki system will index, sign and date your submission automatically.
Seth Horrigan 17:46, 26 February 2009 (PST)
Follow up on my comment for the last class, I read the additional piece from Don Norman, Human-Centered Design Considered Harmful. While I do not agree with Norman on all points, the Activity-Centered Design he suggests makes logical sense. I do not, however, think that Activity-Centered Design should replace Human-Centered Design. On the one hand, he makes the valid point that human-centered design can focus so much on the preferences of a specific group of users that the tool becomes useless to anyone but them, and becomes really poorly designed. Certainly there are times when the design must use their authoritative position based on the design knowledge and say "we will not do this, because that is wrong." Additionally, when the tool must be used by a wider group of people than designers can work with human-centered design could run into problems. That said, in the former case, I have not seen people erring on the side of listening to their users too much; rather, I have seen the inverse. In the latter case, focusing on activity-centered design makes good sense, but even so, you must know about the skills and experience of these main target user group to be able to design the tool to be usable.
Yes, the world is full of non-software tools, and Norman implies that they were developed based on the activity instead of the user. I say that is false. Take a can-opener. Almost all manual ones are designed specifically for people with a certain level of physical hand strength, with a right-dominant hand, and a workable second hand and arm. If you give one of these to someone with one hand, they cannot use it. If you give it to someone with a left-dominant hand, they will have a very difficult time using it; thus, there are hands-free can openers, left-handed can openers, or can openers designed to cut differently, allowing left-handed people to use it well. These are designed with an idea of the activity, but without talking to the actual humans who use it, it will not be a good tool.
Concerning these readings: as I said, I think that Activity-Centered Design is a necessary complement to Human-Centered Design (or possibly a subset of it). Even so, I was not hugely impressed by the first two uses of it. Yes, the toolsets presented could be useful, but I am more interested in seeing the ideas applied to tasks in a natural way. In the first two, they focused so much on "activity" (or on irrelevant implementation details) that the actual activities and their support for them became lost. Tye's paper seemed more appropriate. It still presented a strong focus on applying the idea of activity theory to supporting real activities, but it did so with the focus on supporting the people instead of the focus on demonstrating the theory.
KetrinaYim 13:17, 1 March 2009 (PST)
I, too, was not entirely convinced of the usefulness of the first two activity-centered design tools. After reading the ActivityDesigner paper, I felt it was still unclear as to how ActivityDesigner would support the activities that it was capable of modeling (much less how a swimming fish had anything to do with physical fitness). Bardram's paper was better at convincing me of the potential in activity-based design, though afterward I started wondering what activities other than healthcare routines could the ABC system benefit. Education, perhaps? Or large academic research projects?
Priyanka Reddy 23:59, 1 March 2009 (PST)
Similar to Ketrina, I could see the potential of activity-based design after reading Bardram's paper. Many people are trying to solve the problem with information management in hospitals and I think the ABC system is a good solution for that. To answer Ketrina's question, I think there are many other areas that could benefit from the ABC system, such as the entertainment industry (ie. tv/film production, record companies, etc) and I think the bureacracy could decrease in a lot of organizations if they made use of such a system. Also, I think the ABC system would be a huge benefit to government agencies as they are involved in a lot of mobile and collaborative activities.
I think CAAD is a really interesting tool. While reading it, it made me evaluate my own work flow, and I quickly realized that I have no idea how long I spend on each task that I do when I'm on the computer and I think having more information about that could improve my efficiency. I also thought it was interesting how they find "a disparity between user's mental model of their work and the context structures that CAAD uncovers". When the users brought up that discrepancy, it seemed like the interviewers were suggesting that the users should change their mental model and view their work as more activity and time centered.
Kenrick Kin 02:01, 2 March 2009 (PST)
Interesting. Last week I had a conversation with a friend about how much smarter operating systems could be. We perform a lot of different tasks on the computer and often have many many windows opened. Grouping windows by activities could be a helpful option. If I'm coding, it's likely I'll want to have the terminals, debugger, API (possibly in a web browser) grouped into one activity. Although this could be manually done with virtual desktops, I can certainly see CAAD being useful in learning our activities so that we automatically have ready access to whatever resources are relevant to the task at hand.
Himanshu Sharma 02:09, 2 March 2009 (PST)
Whether the design should be activity-centred or human-centered or both, I think depends on the area of application.
Also, I feel this activity-based design is already being used in many area, even though the designers may not know of the term as such. CAAD paper was easier to relate to personally, particularly the information access scenario. We can certainly benefit from such a system.
David (Tavi) Nathanson 02:48, 2 March 2009 (PST)
@Priyanka I wonder how a tool like CAAD would compare with a tool like RescueTime (RescueTime.com) in allowing people to understand how they allocate their time. RescueTime, like CAAD, uses a logging utility: it logs which applications are being used, and for how long. It then displays various histograms, easily customized by the user, that illustrate how users spend their time. I used it for a few weeks during a summer internship to ensure that I was staying on task.
Regarding logging in general: back in the day, when work was on the computer but play was generally off the computer, logging was not a useful tool for measuring productivity (because you could only log your work). However, the more computerized our lives become, the easier it is to really analyze how effective we are at doing what we need to do (using logging).
Nicholas Kong 03:04, 2 March 2009 (PST)
I personally found the ActivityDesigner paper quite good. Its main purpose was to describe a system that enables rapid prototyping of an activity-based ubicomp application, and I think it succeeded in that regard. However, the design of the application seemed to revolve around the sensing and responses to actions in order to enable or support a higher level activity (e.g., sensing playing badminton to support saying fit). Ketrina, I think the fish display was an aid to keeping users motivated. I thought it was n interesting that the user requested she be able to she her friend's fish; a little healthy competition seems to be an excellent motivator.
A little unrelated, but I also wondered if and how ActivityDesigner's notion of situations fit into the subject/community/object triangle we saw in last week's reading. I suppose situations are governed by laws, but they may be adjacent to the larger activity triangle, or the contexts in which activities live.
Junda Liu 09:59, 2 March 2009 (PST)
@Kenrick, yeah I think an activity based or oriented window management would be very nice for heavy users. I think it will improve efficiency a lot, and keep users focus on current tasks/activity. It seems to me Activity based design is a good complement to human oriented design, not a replacement.
Simon Tan 03:56, 9 March 2009 (PDT)
About CAAD:
The model of studying a user's past behaviours and adjusting a user interface based on that data seems clever but has one known shortcoming - a dynamic workspace is often confusing for users.
Users tend to work better when there is consistency in their workspace; when a display that they are used to suddenly looks different, they often get lost. In an older version of Microsoft Office where the menus would "dynamically hide" commands you don't often use, people would become extremely frustrated as the commands that were left started shifting around. What used to be the "third" command in a menu would suddenly shift up and users would tend to mis-click constantly. This also explains why users hate to have their desks cleaned for them or their desktop icons rearranged.
I was worried that the dynamic nature of the CAAD display would result in the same problem with users, but it may not have been used with enough users or a diverse enough set of users to make it apparent. Alternatively, perhaps the benefits of the system outweighed this potential downside.
Anuj 10:19, 16 March 2009 (PDT)
I am pretty sure there are softwares/applications available in the commercial arena which do things of this sort, but I am less sure on the foundation part of it. I would be really interested in how activity theory can be used to design speech enabled interfaces.
