DITA in Context: At IA Summit 2007

The Information Architecture Summit (http://www.iasummit.org) is a great place to learn from thought leaders in user experience design, interaction design, information architecture, and related areas.

A few people at the conference come from a content background. As an example, Pabini Gabriel-Petit is a user experience architect with a publishing background. Her online magazine http://www.uxmatters.com has a more edited feel than most online publications.

For the most part, though, the attendees are design-oriented. Viewing the user experience as their base, they are interested in the relationship that the user has with the information environment, including but not limited to online user interfaces. To stretch these boundaries, the keynote speaker, Joshua Prince-Ramos, spoke about his career as an architect of buildings. The design process for buildings involves audience analysis, task analysis, and constraint satisfaction, just as the design of user interfaces and documentation do.

The IA Summit is incredibly generous with presentation materials, so you can look at the site to see what was presented. If you can't wait for the presenters to get back to their home bases and post their slides on the official site, you can check out http://www.slideshare.com, where versions of many of the talks may also appear.

I had two interesting hallway conversations about DITA during the conference.

John Allsopp spoke about the need for some standards for describing screen areas. There are a number of dimensions that serve as sources for terminology, such as layout (top, center, bottom) and function (header, content, footer). Accordingly, there is a lot of variability in the classes that are used to associate styles with these areas. He is interested in building a community to discuss how to standardize this terminology. Intriguingly, his paradigm coming in was based on HTML: what value should the class (and id?) attribute(s) have? But there's a case to be made that defining the semantics in an XML language would make it easier to support multiple dimensions. He's aware of the microformats community, but not of the <data> element in DITA. The <data> element might be one level of abstraction too deep for his purpose, but on the other hand, it does offer a great context in which to experiment. This leads to a question for the DITA community:

How would DITA support real screen layout in the HTML class attribute? (And would that support clash with existing DITA presumptions about how to populate the HTML class attribute?)

The second hallway conversation was with an attendee from IBM, Keith Instone. (Thanks to Jennifer Bohmbach for bringing us together.) After a brief conversation about topics and the basic topic types, we came to a question about information centers. When DITA topics are delivered in an Information Center as they are currently constructed, the intended interaction is that the user would switch from the application to the Information Center in order to explore supporting content. Since the DITA architecture is rich in means to integrate chunks of information, this leads to a potentially fruitful follow-on discussion.

How can the DITA architecture and the architecture of the contexts in which DITA is used be enhanced and extended to permit better integration of DITA content into user interfaces?

As Michael Priestly noted in his blog entry MP: CMS/DITA 2007 Day 1, the Total Information Experience (TIE) initiative at IBM is pointed in a similar direction, with a somewhat broader scope. It will be interesting to see how these communities, experience design and content architecture, influence one another as they become aware of each others' issues and approaches.

-- Bruce Esrig

Hi Bruce,

I have been a member of IA Institute since 2003 and went to my first IA Summit in 2004. 

It was there that Bob Boiko, Tony Byrne, Ann Rockley, and I, with advice fromPeter Morville and  Lou Rosenfeld, decided to start the CM Professionals community of practice, modelled on IAI.

I agree that IA's are very generous with their knowledge. The IAI website has lots of great resources, for example.

But I find that their primary interest is in navigation among documents, things like findability and usability. Their library science background shows.

They have little interest in structuring the content inside a document. They have mostly never heard about Information Types like those  in DITA.

I wrote an EContent column about this.  IA's > CA's and DA's.

I just looked over the IA Summit 2007 program, especially the workshop "Designing with Structured Data," and see they have tried to correct this negative impression of mine. That may be why you were there?


Bob Doyle

Hi Bob,

I would say that I'm finding the IA Summit interesting because of the degree to which the IA community is engaged in interaction design. This has profound implications for our priorities when we work on content.

If you define IA as "information architecture", and interpret the phrase word by word, it is fair to say that the IA community primarily discusses information looking from the outside in. The addressable unit is anything you can get to via a URL. Tagging imposes a user's notion of information types from the outside.

Taking that definition of IA, screen layouts are the first place that information types seem to show up. You provide types to the parts of the screen, and do different things in the different regions. In portal world and gadget world, the behavior is associated with the portlets or gadgets, rather than with fixed positions on the screen.

But I'm attracted to a broader definition of IA, in which the users are actively interacting with content. (Granted, when a CMS is used to support information delivery, the users are live there too.) To reframe information types so that they apply to live interactions, you have to be prepared to consider types for user behaviors.

This leads to interaction types, which are types for the behaviors that occur while users are performing their tasks. Interaction types are jazzier than information types, although what you can do with them isn't all that different:

  • Assign a type to a behavior
  • Standardize the functionality and presentation of an interaction based on its type
  • Organize controls in predictable patterns according to type
  • Search for functionality based on the interaction type.

Just substitute "information" for "control / behavior / interaction / functionality", and it all reduces to things we do with information types.

So the IA Summit is very interesting to me because of the struggle that I see people engaged in to deliver information to one another. There's a genuine feeling of excitement as user experience designers work to break down the barriers that interfere with people accessing the information they seek. The explosion of community-oriented features and immersive experiences provide new contexts in which people encounter information. These contexts don't render content obsolete, but they radically change people's priorities about how they wish to solve problems.

If we are delivering content in order to support people in solving problems, it makes sense to switch points of view. Start by looking at the problems people are trying to solve, and the contexts that they seek out in creating their problem-solving environments. In those environments, what is the role of authored, edited content? We need to view user problem solving as a whole in order to see what contexts to provide, and what information to provide in those contexts.

Another aspect of user interaction is the emotional attraction that people feel to sites and the judgements that they make about both sites and their creators. To explain this area, it is necessary to understand not just user goals, but user motivation, i.e. how the priorities among user goals change with context and time.

Although the IA Summit was founded in large part by library science people, these issues of interaction and emotion have entered the discussion, and are carrying the field of information architecture way beyond information retrieval. In support of that perspective, it was interesting to hear people talk about how they explain what they are doing in corporate terms. User experience design was a very common answer this year.

Bruce Esrig
XML.org Focus Areas: BPEL | DITA | ebXML | IDtrust | OpenDocument | SAML | UBL | UDDI
OASIS sites: OASIS | Cover Pages | XML.org | AMQP | CGM Open | eGov | Emergency | IDtrust | LegalXML | Open CSA | OSLC | WS-I