I follow the blog of Irving Wladawsky-Berger of IBM Research (hereinafter "W-B") because his insights are often more profound than one generally finds in blog posts. A great example is his recent post entitled The Importance of Human-Oriented Design to Avoid Drowning in a Sea of Complexity. This post talks about usability in very concise, practical terms grounded in concrete examples. Where a usability pundit like Bruce Tognazzini ("Tog") issues lists of First Principles of Interaction Design, W-B's post gives you a deeper understanding of why usable applications are usable . This helps you understand when Tog's rules apply and why it is sometimes useful to violate them.
Don't get me wrong - Tog's list hands on my office wall, where I see it every day. I bounce ideas off it as if it were another person in the room. Nonetheless, his First Principles have their limits, and they are not apparent from reading his essay on the subject. Understanding what W-B says in his post will give you the intellectual grounding you need to develop an innate understanding of Tog's First Principles, if not to derive them yourself. Here's some comparisons that make me feel the way I do.
Tog:
Effective interfaces are visually apparent and forgiving, instilling in their users a sense of control. Users quickly see the breadth of their options, grasp how to achieve their goals, and do their work.
Effective interfaces do not concern the user with the inner workings of the system. Work is carefully and continuously saved, with full option for the user to undo any activity at any time.
Effective applications and services perform a maximum of work, while requiring a minimum of information from users.
When I visit my online banking sites, this describes exactly what I want and need in an interface. These three edicts elegantly summarize the essence of his First Principles.
W-B starts from a different but complementary point of view:
... successful IT applications – those that appeal to large numbers of people -- ... appeal to us because they are intuitive and thus easy to learn and use. What makes applications intuitive is the fact that they are designed around objects from the real world that people are generally familiar with, and thus we can bring our real world knowledge and intuition to bear on the applications - they have a point of connection, as it were.
The applications are then developed for the virtual world of computers, including the realistic simulations of the physical world objects on which they are based, along with added features which good designers will also make as intuitive as possible. Since our interactions with the physical world are so visual in nature, it is not surprising that the more visual an application, the more intuitive it is likely to feel.
Perfect - it's easy to see how Tog's adages apply to my online banking site, and a host of other applications I use, on the Web and otherwise, and to a variety of objects with a designed interface. However, W-B goes into more depth on what it means for an application to be intuitive, with some concrete examples. He says that word processing is about typewriters and documents, and such programs' WYSIWYG (What You See Is What You Get) nature allows the representations of documents to look as close as possible to the document as it will appear in real life when completed.
| Application | Domain objects represented | What makes it intuitive |
| Word Processing | typewriters and documents | The representations of documents to look as close as possible to the document as it will appear in real life when completed. |
| Web browser | A page from a newspaper, magazine or book, or a form, albeit significantly enriched with all kinds of added features like hyperlinks | The highly visual and realistic presentation of content and form-based applications in browsers has played a key role in their universal appeal. |
| Engineering applications | The object being designed, e.g., a car, airplane, building, bridge, perfume bottle, or article of clothing. | (CAD) applications build on the notion of the drafting tables used by engineers, architects and designers of all sorts to sketch out their designs in increasing levels of detail. Good CAD applications are intrinsically visual, focusing on the look, properties, style and performance of the object being designed. |
| Scentific applications | Any entity, event, or process of scientific interest, e.g., a chemical reaction, a hurricane, or brain structure or functioning. | Increasingly vast amounts of data are best understood when visually represented, whether highly realistic (e.g., an animated overhead view of a hurricane), augmented (e.g., functional MRI), or metaphorical (network analysis visualizations of the Internet). |
Hmmm... some of these have very different usability criteria than my bank's personal banking application. Word processors and Web browsers still fit the bill, for the most part. But CAD applications, for example, are notoriously complex and hard to use because the only people who use them at all are people who must become expert users to use it to any productive end at all. There is no reason to invest in UI features to support the "casual user"; there basically aren't any, and screen real estate is too precious to waste very much of it on handholding devices. Moreover, the companies may insist on training end users, to reduce the support burden, and provide bootcamp-style classes to bring users up to speed. They can charge thousands of dollars per seat for their software and thousands more for the training because the ROI of such an application is almost incalculable to an automotive or aerospace manufacturer, when employed by an appropriately managed expert user.
Scientific applications are often relatively obscure in terms of usability because they were created by domain experts without any background in software design, much less ergonomics and human factors. Their peers grab their software and run with it, warts and all, because all they care about is the functionality. They also may enjoy toying with the application and reading the code in order to figure it out, but functionality is key - nobody wins a Nobel Prize by choosing the easiest-to-use analytical tool.
Still, most of Tog's principles apply in CAD and scientific applications. They have to meet a minimum bar in terms of being "siually apparent and forgiving," for example, no matter how useful they may be, and for the most part they benefit from hiding the inner workings from end users - or at least shielding them from the necessity of dealing with the inner workings.
There is, however, a paradigm shift underway in the realm of applications, a sea change that turns Tog's edicts on their heads. W-B was one of the first people to clue me in to this new realm, and he reiterates it here: serious games.
The leading edge of intuitive application design is now taking place in the world of video games. The whole object of these highly interactive games, whether developed for single game players or for online massively multiplayer platforms, is to enable the players to make decisions very quickly. Thus, their interfaces have to be truly intuitive. Video and online games are the ultimate WYSIWYG applications.
Applications like Doom or Myst, for example, could never have succeeded if they were "visually apparent and forgiving", or if the users could "quickly see the breadth of their options". What fun would that be? Moreover, the inner workings of Doom were deliberately exposed so other people could leverage them to create ever-richer and more varied games based on the Doom Engine.
Now think about something like Second Life $$$$.
Ultimately, W-B and Tog have a lot of common ground. Tog would, I'd bet, heartily agree with this W-B statement about the limits of technology:
...technology advances, while absolutely necessary, are far from sufficient. The hardest challenge of all, but also the most promising opportunity for breakthrough innovation, is to understand how people best perceive and interact with applications – e.g., what makes an application feel intuitive, industry by industry and process by process. How can we design hospital management systems, supply chains for retail stores or HR applications for a bank that are truly human-friendly to all the various people who use them?
This work will require a lot of talent and a variety of skills. It is truly one of the areas in the IT industry most in need of innovation, but also one full of opportunities for those who take it on.
In my view, the most important skills will be some of the oldest and most basic human emotions: empathy and compassion. A good computer program in any domain relieves pain and creates joy. These skills aren't taught in engineering schools, and only indirectly and non-verbally in the HCI classes I've taken. Then again, I'm not sure they can be taught; they are the fruits of the maturation of one's character.
Comments