This is a debate between two streams of user interface research: direct manipulation of interface objects versus delegation of tasks to software agents.
Shneiderman believes an interface should be evaluated by the performance of its users; this begs the creation of an environment where the users can “comprehend the display”, “feel in control”, and “take responsibility for their actions”.
Maes concerns about the ever increasing complexity of the computing world; this has brought forward the ideas that much of what people do can be delegated to software agents, thus easing users’ burden in carrying out these somewhat difficult or tedious tasks on their own.
- The overview is the most important. It gives users a sens of context, of what to look at – the big picrture. Then they zoom in on what they want, filter out what they don’t want, and finally go for details-on-demand. My claim is that this gives users the feeling of being in control and therefore they can be responsible for the decisions they make.
- Software agents:
- A software agent knows the individual users’ habits, preferences, and interests;
- A software agent is proactive. It can take initiative because it knows what your interests are;
- A third difference with current software is that software agents are more long-lived;
- Software agents are adaptive in that they track the user’s interests as they change over time.
- Misconceptions about software agents:
- Agents are not an alternative for direct manipulation;
- Some people think that agents are necessarily personified or anthropomorphized;
- Agents necessarily rely on traditional AI techniques, like knowledge representation and inferencing.
- It’s as if you had someone else looking over your shoulder as you are using the application, noticing some of your preferences and habits and then offering to automate some of the tasks for you.
- In particular, we need to take care of these two issues: understanding and control.
- Understanding means that the agent0user collaboration can only be successful if the user can understand and trust the agent;
- Control means that users must be able to turn over control of tasks to agents but users must never feel out of control.
- User understanding is central, and user control is vital for people to be successful.
- Two levels:
- The user interface level, which users want to be predictable and controllable;
- The level below the table, where there may be some interesting algorithms such as collaborative filtering.
- I think what annoys me the most about the devotees of speech, is their failure to take in the scientific evidence that speaking commands is cognitively more demanding than pointing. Speech uses your short-term memory and working memory. By contrast, hand-eye coordination can be conducted in parallel with problem solving by another part of your brain and therefore does not degrade your performance as much as speaking.
- Please, please, please do your studies – whether they are controlled scientific experiments, usability studies, or simply observations, and get past the wishful thinking and be a scientist and report on real users doing real tasks with these systems.
Direct manipulation is not everything. Reading this debate dawns on me that it could triggers so much interesting thought simply by comparing our domain with others’. It forces us to ponder several questions: what exactly is direct manipulation? What does direct manipulation do compared to software agents? What are the advantages can disadvantages of direct manipulation? What will become the future of direct manipulation and software agents – one wins over the other, in symbiosis, or blend into a new concept?