brownorama’s bookmarks, posted 2005-03-31
brownorama’s bookmarks, posted 2005-03-31
On Monday night, I attended a the local UPA chapter seminar called Task-Free Usability Testing. During the session, three panelists described their experiences with a “new” approach to conducting usability tests. All the panelists admitted to having employed these techniques to a lesser extent previously — which is why I qualified “new” — but for each this was the first time they had built a structured test around it.
The technique boiled down to keeping the test open-ended, not forcing users through pre-determined scenarios or tasks. Instead, users were invited to explore the site on their own and comment. Each panelist presented a slightly different version of this theme.
Lisa Battle spoke first, describing her experience working on the Social Security Administration’s web site. Because the project did not include time for sufficent user research, she compressed requirements-gathering with a scheduled baseline usability test. She asked participants to talk about different scenarios in their lives relating to Social Security, and captured these scenarios on cards. When it was time to test the site, she pulled these cards out and asked participants to imagine themselves in these scenarios.
Describing her work for National Cancer Institute, Christy Mylks spoke next. In her test, participants first spent time looking at the existing Web site, then met together in a group to discuss. After the group discussion, they looked at a prototype for a redesign and Christy’s team conducted one-on-one testing. In this case, the group discussion helped elicit and elaborate on some of the requirements — like how breast cancer patients don’t like registering for web sites because they have had to give up so much information already.
Finally, Lynn Baumeister explained how her work for Genome.gov also involved more open-ended testing. She recruited through her personal network and visited particpants in their offices, allowing them to explore the site on their own. She noted one side effect of this approach was that participants identified important supplemental information they would want to support other content on the site. Lynn referred to this as the “bubble” of information.
The major disadvantage to the open-ended approach is lack of “hard data.” Because there is little structure to these tests, there is no way to capture accurate numbers. To me, this is hardly a disadvantage. Numbers attached to usability testing always seem like a bit of a reach anyway.
Take-away #1: Recruiting is crucial.
These open-ended tests demanded interested participants, motivated participants. With scenario-driven usability testing, it is perhaps easy to settle for participants whose interest is not necessarily specific to what you’re testing. Although the task-free testing could save you time in planning, I would merely shift that time to recruiting activities.
When you surround yourself with ideal users, it is easier for you to become their surrogate during the design process. Jeff Veen described this as the Method Acting approach to design, an idea that makes a lot of sense to me. A good designer becomes the user — gets so far inside the head of user as to be almost indistinguishable.
This is not to say we can eliminate users from the design process altogether. Instead, it makes the interactions with users more meaningful: help me become one of you. The open-ended testing techniques discussed on the panel is one very good way of doing this.
Take-away #2: Long live informality!
At the beginning of the panel, Christy indicated that some of the thinking around this open-ended technique started at the last UPA national conference. At this event, the backlash against lab-based testing begin: people started questioning the value of taking people out of their “natural environments” to see how software or websites performed. The emphasis should be shifted, say the backlashers, to contextual inquiry.
This smacks of the eternal battle between Jedi and Sith, Elves and Orcs, formality and informality. On the one hand, we have the formal approach — driven by “experts,” and highly structured — which can lead to hard data and solidly defensible conclusions. On the other hand, approaches with less structure and open-ended goals give practicioners greater flexibility. Usability isn’t the only field experiencing tension between formal and informal. The (somewhat) famous debate between folksonomies and taxonomies is also a discussion about formality.
Some ideas to think about:
brownorama’s bookmarks, posted 2005-03-28
Caveat: I write this without having done my homework. Aside from attending PeterMe’s talk at the summit and perusing some of his follow-up posts, I haven’t read the canon on genre. If my thoughts sound ridiculous, it’s because they are.
What is a genre?
Quoted by PeterMe from Orlikowski and Yates:
“a distinctive type of communicative action, characterized by a socially recognized communicative purpose and common aspects of form�
My interpretation: a genre is cue that sets consumers’ expectations. Different genres set expectations differently: some focus on form while others focus on topic. There may be different kinds of genres, as illustrated by the difference between “comedy” and “textbook.” Both are genres in that they both set user expectations about the form and scope, but they’re clearly two different animals.
Genre as a graded category
There are some genres that are better examples of a genre than others. If the main criteria of the genre prototype is that a genre sets expectations about form, content, topic and scope, among other things, then non-central genres are those that do not do all these things.
Genres as graded categories
Each genre is itself a graded category. There are some comedies that are better examples of a comedy than others. What’s interesting is that since genre deals with expectations, it’s possible for something to be in a genre without it exhibiting any of the expected characteristics. Consider comedies that aren’t funny. They are considered to be in the comedy genre, and yet entire audiences will barely giggle. We don’t seek to re-categorize it. Instead, we just say it’s bad. As another example, you could say “In my IA class, our textbook was the Polar Bear book.” Technically, Polar Bear is not a textbook, and yet it still falls into that genre, though no where near the prototype.
Genres as frames
Framing is an aspect of language, where speakers imply a set of background assumptions. In political discourse, for example, the phrase “tax relief” frames the discussion about taxation in terms of a burden. Because genres deal with expectations they can make excellent frames. In an information space about a given discrete topic, links to different genres will frame the “conversation” about the topic in different ways. A history book about Washington, DC frames the topic differently than the Washingtonian.
Genre as metonym
A metonym (as I understand it) is a part of a concept that can stand in for the whole concept. Dates and locations are great examples. When you hear “Today the whitehouse announced…” on the news you know that the location “whitehouse” actually refers to the institution that resides there. We’ll often use genres as placeholders for content. When you point at a map and say, “That’s a one-way street,” you don’t mean the lines on the map itself.
Genres as blends
Genres allow us to blend. Blending is a mental activity — an automatic imaginative response to new ideas and helps us make sense of them. When creating blends, we’re smart enough to project only the relevant parts from the basic concepts. With a map of Montreal, for example, we’re blending the geographical representation with the city, but only certain parts of Montreal are projected into the blend. Likewise, a map of the ocean floor is still a map, but requires a different sort of projection. When we imagine these maps we’re creating mental blends: form with content.
Genres in digital space
James has been doing some interesting thinking on different views of a genre. In digital space, since we have some flexibility over the presentation, we can reduce a piece of content to its barest form without “losing the genre.” James was calling this the “minimum shape” or “minimum appearance.” IAs face a difficult challenge: how much can you change the shape of something before it loses its genre? A press release represented in a list of press releases is still of that genre.
In our discussion, James and I explored how genre is applied to re-used content. Amazon product information gets re-used throughout the site — sometimes as a full product page, sometimes as a link on the home page, sometimes as a link on your wishlist. The question is which is the genre: the core product information, or each of the representations?
This question brings us to:
The controversy of genres
In his presentation, PeterMe said that in genres, “presentation (form), content, and purpose are intertwined”. This did not get the response from the audience I thought it would. I mean, he basically said that the idea of splitting content from presentation is not all its cracked up to be: the content is is the presentation, and vice versa.
From an expectations point of view, the form and the content are inseparable: content must be understood through form. To take data out of a table is to remove all meaning from it. To take “content” out of a book or Web site is to remove all context. By saying something is a textbook, we are making a claim about its form. We can not then turn around and “repurpose” textbook content into PowerPoint content: the genre places demands on the form of the content at all levels, even down to the style of writing.
Despite the proliferation of the work “folksonomy” and my respect for fellow DC information architect Thomas Vander Wal, I’m partial to the term “freetagging.” Freetagging — applying freely-chosen words as labels to objects — is the activity that leads to a folksonomy. When you type tags into del.icio.us, you’re freetagging.
My preference for freetagging comes from looking at what makes the activity valuable: the process itself. A folksonomy is an ever-growing vocabulary that is internally inconsistent (out of necessity). The value is not in its content or structure, but in how its generated. Compare this with the aptly named “controlled vocabulary,” which is labeled based on its value: there is a limited set of terms you can use when categorizing stuff.
The value of a controlled vocabulary is the vocabulary itself, not the process for getting there. This is not to diminish the work required to create one — it is very labor intensive. The point is to focus on what’s important: a controlled vocabulary is a tool. On the other hand, a folksonomy is not a tool — it’s the by-product of a tool, the capability to apply personally-chosen labels to objects.
Is this what IA has come to? Nit-picking jargon? Can friends find nothing else to discuss over coffee?
Pretty much. But I’ve been thinking a lot about an article by Bruce Sterling in the latest Wired, in which he compares (informal) folksonomies with (formal) taxonomies. It’s apples and oranges. He *should* be contrasting folksonomies with controlled vocabularies. Here’s my take:
| folksonomies | controlled vocabularies |
|---|---|
| theoretically infinite scope | limited scope or domain |
| no formal vetting process for changes | changes require formal vetting process |
| constantly expanding | periodically expands and contracts |
| accidental relationships between terms (when two terms are used to tag the same object) | explicit relationships between terms |
| flat | structured |
| employed in wayfinding (the folksonomy itself is never consulted to determine appropriate terms — but that is both its strength and its weakness) | employed in both wayfinding and classification |
| proprietary (although in most applications everyone’s tags are exposed to everyone else, a person’s list of tags is unique to him or her) | shared |
| meaningful to individual | meaningful to everyone |
Sidebar: this last one requires a bit of explanation. As an example, consider Flickr’s image library compared to a stock photography library. The images in Flickr are tagged with terms that are most meaningful to the author, so a picture of my wife I’ll tag “Sarah.” If someone were to come to Flickr and type “woman” into the search field, my picture would not show up in the results. Compare to a stock photography site, which employs a controlled vocabulary to ensure that such searches reveal something useful.
Sterling, for a futurist, is incredibly short-sighted in this article. Although he doesn’t imply that information architects are completely useless, he does suggest that with advancing technology and contributions from the public, IAs will need to focus on formal information domains — like law and medicine. I don’t think he understands the role of IA in all domains. For example, he lauds the powerful search of Google: people will be able to find whatever they want. My questions for Sterling:
Who devises the search alogrithms?
Who designs the interface for presenting results?
Who develops the mechanism for storing metadata?
These aren’t just technology questions — these are IA questions!
brownorama’s bookmarks, posted 2005-03-25
A recent story in NYTimes > Books describes the evolution of a genre: the memoir. Some choice quotes illustrate that the article author sees evidence of loosening the criteria:
[Ulysses S. Grant] produced a classic memoir, as the genre was then understood: important events related by a great man who shaped them.
But the genre has become so inclusive that it’s almost impossible to imagine which life experiences do not qualify as memoir material.
In apparent exasperation, the author then goes about creating a “memoir taxonomy,” supposedly to “impose some sort of order on this sprawling genre,” with categories like:
There’s clearly some Lakoff-like categorization going on here, because the author bemoans one memoir that “has a little bit of everything: military combat, a spiritual awakening and lots of prescription drugs.”
The article takes a more serious turn when he talks about humanity’s need to create narratives about themselves, pointing to several examples of schools that teach biography writing, of informal reading/writing groups sprouting up all over the country, a proliferation of ghost writers for busy executives. It’s almost as if it were a natural tendency, this talking about ourselves thing.
I see the memoir genre changing for different reasons. Like a meteor hitting the earth, blogs changed the landscape for all written genres, particular ones like memoir, which is defined as a personal story. Memoirs, as far as genres go, may not be dinosaurs, but they do need to find their identity in the new digital landscape.
Believing that genres can, should, and will remain static is to deny the fundamental cultural aspect of genres. They are cues invented by groups to set expectations regarding content. Cave paintings were probably a meaningful genre for early humans, just like certain religious rituals and newspapers are for other groups. As the cultural landscape changes so, too, will genres.
After playing with three different whole-house audio systems, we’ve decided to go with the sneakernet — carrying the iPod around the house. The iPod offers several advantages:
There are some cool add-ons for the iPod, like a remote control. Another promising device allows you to send music from the iPod wirelessly to a pair of headphones. (We’re going to look into plugging the device into a stereo.)
Shared bookmarks for del.icio.us user brownorama on 2005-03-23
