In this week’s seminar we left the discussion of information and began looking at the theory of design writ large, with a few focused readings, and a lecture from Professor Findlater. One of the key things I took from the lecture was the distinction between User Experience and Usability. The usability of an object speaks to its functionality, and how easy it is for people to use it. User Experience on the other hand is more of an affective measure of how users perceive or think about the device, which may not always be congruent with it’s usability. It’s an interesting distinction, which I wasn’t really conscious of before.
Speaking of distinctions, we spent a fair bit of time talking about the first chapter in Don Norman’s The Design of Everyday Things. Norman is well known for popularizing the term affordance in the HCI community. He borrowed the term from the psychologist James Gibson. We read from Norman’s revised edition from 2013–the original was published in 1988. In chapter 1 it seems that Norman has a bit of an axe to grind because of how affordance had been used in the literature to denote a relation between an object and an organism, as well as a perceived affordance or what he now calls a signifier. This might seem like splitting hairs a bit, and perhaps a bit quixotic after the term affordance has been out in the wild for 25 years–but for me it made some sense. I know everyone didn’t but I appreciated his sometimes acerbic tone, especially when he was berating General Electric for its continuously flawed refrigerator controls.
A good example came up in class of a student who recently moved and needed to buy a couch. She is tall, and often likes sleeping on a couch. So she was looking for a couch that would be easy to sleep on. Specifically she wanted a couch with arm rests that could also serve as pillows, because she is tall. For example compare these two couches:
The second couch affords being used for sleeping by a tall person. The affordance is a specific relation between the tall person and the couch, not a specific feature of the couch. The distinction that Norman is making here is that the affordance is different from the perception of the affordance, or signifier, which in this case is the type of arm rest.
Being able to distinguish between the relation between the object and the person, and the sign of the relationship seems important–particularly in cases where the affordance exists but is not known, or when it appears that the affordance exists, but it actually does not. I’m thinking of controls in a user interface that appear to do one thing, but do something else, or do that expected thing as well as something unexpected. I can see why HCI people would have reacted negatively to Norman telling them they were using the term wrong. But since I’m new to the field I don’t have that baggage I guess :-)
A few other things that came up in discussion that I will note down here to follow up on at some point are:
- A Hierarchy of Complexity: Bill Moggridge (founded Ideo and was Director at the Cooper-Hewitt)
- Eight Golden Rules of Interface Design: Ben Schneiderman (started the Human-Computer Interaction Lab (HCIL) at UMD.
- The Humane Interface: Jef Raskin
In the second part of the class we had a writing workshop where we discussed writing we liked, writing strategies (since we’ll be doing lots of that in the PhD program), as well as ways to get started writing when you are stuck.
We all brought in examples of writing we liked, and talked briefly about them. I brought in Jackson & Barbrow (2015) as an example of what I thought was a well written paper. I like it because it is strongly grounded in enthographic research (admittedly something I want to learn more about), and discusses a topic area that I am interested in: ecology, measurement and standards. I thought the paper did a good job of studying behavior around standards at different scales: the individual team of researchers out in the field, and the large scale national collection of data. Jackson included numerous quotes from individuals he observed during the study, which added additional authentic voices to the paper. He also quoted Borges in an useful/artful way. I thought it was also very interesting how standards were presented as things that we should consider as a design factor in Human Computer Interaction. The very idea that seemingly invisible and dull things like standards (how they live, and what they omit) could be useful in design is a fascinating idea to me. I liked the illustration that while measuring the environment seems like a precise science, it has a human/social component to it that is extremely important. Finally, I’ll admit it: I brought the paper because it won an honorable mention best paper award at CHI – and I’m a fan of Jackson’s work.
After we all talked about our favorite papers we discussed writing strategies or techniques to be aware of:
- illustrations are important, captions are important
- formulas as visualization
- subheadings to make things easier for the reader
- start a section with the main point, so people can find their way back
- template, protocol for experimentation: group study
- goal is sometimes to help people replicate
- need a clear method: this is a must
- need to be able to justify the decision
- supplementary information (datasets, interview transcripts, coding, etc)
- talk about failure
- made clear real world implications
- in first paragraph that the research is important
- reflection on findings
Ways to get started
We also talked about ways to get started writing when it is difficult to get going:
- look for similar papers: content & venue
- keep track of good examples of papers
- may find some that work as templates
- talk about the paper, get feedback all along the way ; talk to co-authors
- talk to the people that you are citing
- start with what’s easiest: sometimes lit review, or methods, depending
- what is the story: what do you want people to remember, how do they get highlighted ; the turns
- where do you like writing?
- find your voice, style, etc.
- give yourself time
- the time of day is important
- stretches of time are important (take breaks)
- be clear and concise
- use evidence
Here are my random notes I took while doing the readings for this week. They haven’t been cleaned up much. But I thought I’d keep them here just in case I need to remember something later.
When doing the readings we were asked to pay particular attention to these questions:
- What problem or research question does it address? Is that an important problem/question?
- What is the research contribution?
- Is the overall research approach used (e.g., lit survey, interview study, experiment) appropriate?
- Are there any threats to the validity of the research? For example, the sampling method was for the researcher to ask all their friends to participate (let’s hope we don’t see this!).
- Is the research properly grounded in past work?
- Are there any presentation issues?
Norman (2013), Chapter 1
- Psychopathology: the scientific study of mental disorders, and their causes (potentially in the environment)
- Is it possible to learn how a product is supposed to be used?
- “All artificial things are designed.”
- Types: Industrial, Interaction, Experience (last two seem to blend a bit)
- Rules followed by a machine are only know by its designers – and sometimes not even then (Nick’s work on Algorithmic Accountability)
- The importance of blaming machines for the problems, not the person. Human centered.
- “If only people would read the instructions everything would be all right.” Haha: RTFM.
- Wow, he studied the usability of a nuclear reactor and Three Mile Island. No pressure!
- Need to design for the way people are not the way you (the designer) want them to be.
- In Human Centered Design (HCD) focus on breakdown, not on when things are going right. Winograd & Flores (Computers and Cognition).
- HCD cuts across the three types of design.
- Affordances: relationship between a physical object and a person (or agent). An affordance is not a property of the object, it is a relationship.
- J J Gibson came up with the word.
- Some are perceivable, and some are invisible.
- Norman introduces: affordances, signinfiers, constraints, mappings, feedback, and conceptual model.
- Signifiers are signals that draw attention to affordances, or make them visible.
- They can be accidental or unintentional. Markers of use (a path through the woods)
- Some signifiers are the perceived affordances, useful in cases where the affordance isn’t easily perceived.
- They must be perceivable, or else they are failing to function.
- Some perceived affordances (signifiers) can appear real, but are not. (used in games, and other places to illustrate constraints?)
- If signifiers are signals for affordances, I guess affordances are the signified?
- Mappings: a relationship between the control and the result.
- some are natural, some feel natural (but in fact are not)
- Feedback communicates the results of an action: there can be too little feedback, and too much feedback (backseat driver).
- Conceptual models are explanations of how something works. Often they are simplified, and there are more than one. They are often inferred from using a device.
- System Image: the bundle of all these things – akin to an actual brand?
- Hardest part of design is getting people to abandon their disciplinary viewpoint and to think about the person who is using the device (antidisciplinarity)
Crystal & Ellington (2004)
- This paper is a review of the literature on the topic of task analysis, with a view to the future, so not so much a formal research study really.
- It seems like a pretty thorough review, but wondering if it is a regurgitation of Kuutti and Bannon (1991) that is mentioned in the conclusion. Although I guess Crystal reviews content past 1991, so maybe it’s an update of sorts?
- It’s kind of amusing how they use their own task analysis breakdown to present the illustration of the field. I guess this would be a conceptual model (CTA)?
- My impression was that more could’ve been done to define notion of usability which is thrown in at the end.
- I liked how they wanted to compose and integrate the research on task-analysis, and not validate a particular viewpoint, but show the options that are available, and suggest cross-pollination between the models. Not sure if this is inter-disciplinary or not.
Oulasvirta, Kurvinen, & Kankainen (2003)
- conclusions are inconclusive
- hypotheses are introduced too late
- good grounding in the literature
- they seemed to have tow different variables at play: documentation provided and environment
- admittedly limited analysis of the design documents
This study is built upon a useful foundation of existing research on bodystorming, and seems to provide a useful introduction to the concept. It also usefully highlights through concrete examples how bodystorming is different from brainstorming. The goal of the study seems to be to be explore two hypotheses, that are introduced at the end of the paper (I think they should’ve been included earlier):
- to speed up the design process
- to increase awareness of contextual factors in design
The authors mentioned a third hypothesis, which was to evaluate whether bodystorming on site provided immediate feedback about design ideas. But to me this seemed like a very minor variation on the speed of the design process.
They attempted to study these questions by analyzing the designs generated by 4 different case studies where bodystorming was used and a more traditional brainstorming case study. The setting of the bodystorming was varied: on site, similar site, office space, office space with acting. It also seemed like different types of documentation (user stories and design questions) were used in each scenario. This seemed to be changing more than one variable, and complicating the ability to draw conclusions. The authors mention that the results were somewhat complicated because designs from one of the bodystorming sessions seemed to inform other sessions. This was strange because they mention elsewhere that the case studies included different participants; but they couldn’t be all different if this sort of learning took place?
The findings themselves were inconclusive, and admittedly somewhat shallow. Although some of the anecdotes regarding site accessibility, level of exhaustion, inspiration and memorability seem like they would be fruitful to explore in a more controlled manner. It felt like this study was trying to do an experiment, but really did a much better job of presenting the ideas of bodystorming in the context of the literature, and providing a useful set of case studies to delineate how it could be used.
Crystal, A., & Ellington, B. (2004). Task analysis and human-computer interaction: Approaches, techniques, and levels of analysis. AMCIS 2004 Proceedings, 391.
Jackson, S. J., & Barbrow, S. (2015). Standards and/as innovation: Protocols, creativity, and interactive systems development in ecology. In. CHI; Association of Computing Machinery. Retrieved from http://sjackson.infosci.cornell.edu/Jackson&Barbrow_StandardsandasInnovation(CHI2015).pdf
Norman, D. A. (2013). The design of everyday things: Revised and expanded edition. Basic books.
Oulasvirta, A., Kurvinen, E., & Kankainen, T. (2003). Understanding contexts by being there: Case studies in bodystorming. Personal and Ubiquitous Computing, 7(2), 125–134.