Magic Inside vs the Coherent Universe

Reflections on the evolution of user interfaces

Have you been hanging around common people lately, watching how they use their computer systems? They often seem TERRIFIED from using their computers to their full potential. They learn how to perform their usual four or five tasks, and any deviation of their known flow (such as an error dialog) often means that they won't be able to finish the task.

Why is this so? If you ask them, what they will tell you will fit into this two main reasons: because of the very real possibility to break the system, and because most actions are overly complex even for simple tasks such as syncing their music folder with their mp3 player.

That's the reason why task-based flow environments, such those found in the Post-WIMP devices such as with Android and iOS, work much better than object-oriented environments for these users; self-contained tasks offer all the information users need to achieve their goal, while in desktop environments the user must have a thorough understanding of all the virtual objects and their relationships before anything cat get done at all.






Task-based

  



Object-based

Users simply don't have the background nor the time needed to learn all those concepts, since the objects are designed in a way that it's hard to make sense of. I call the first task-based style the "Magic Inside" model, and the second one the "Coherent Universe". People enjoy pressing buttons and having the desired result appear, as long as the right magic is achieved and no unexpected things happen. On the other hand, making sense of a whole universe is likely to be overwhelming, though it allows for much more flexibility. These styles are by no means new, they are  design forces that have appeared and been discussed in traditional interfaces. Usually good designs strike a good balance in the middle.

Legacy controls, and the future

The WIMP widgets certainly were extremely well designed and flexible for their time, and that's why they could have survived all those years at all; and if you think of it, they were at first quite task-oriented in the way that I explained above. But they've largely overexceeded their useful life for many purposes, and they're not well adapted to the needs of modern applications (how many menu bars do you see on Twitter, Facebook, Google or Wikipedia?). They've not been replaced because we haven't found an optimal replacement yet, and because the HUGE investment in existing tools (widget libraries, integrated development environments, and peripherals) supporting their model. Those make for a strong conservative force. (Yup, developers are extremely conservative people when it refers to their profession tools).

Certainly some gestural and natural interactions are gimmicky when applied to traditional tasks. The trick is that they will allow for new kinds of tasks, those that have been held back by the dominant desktop metaphor - primarily, sharing information between people on the same physical and/or virtual room, and working on analyzing and editing huge data sets.

Design of desktop environments for the common user

System builders ofen try to create a flexible interface by cramming it with configuration options, in the hope that naive users will find a simple default configuration that works well, and expert users will be able to set up the interface to match their needs.

In my view there are two kinds of configuration options. The first one allows the user to create new workflows (adding widgets to panels and the desktop, installing apps from the market...) and is useful for adding powerful options to the system, and the second one (far too common in Free and Open Source software) is used to get rid of design decisions that get in the way of the preferred workflow.

New FOSS desktops like Gnome 3 and Unity are attempts to simplify and reduce the need for configuration options. In its effort to get rid of the second kind of configuration, these projects seem to have made the mistake to also remove the first one. That doesn't mean that the underlying design is the wrong one, centered around a task model that can be expanded to accomodate new tasks; it's just that they have made ther system simpler than it should be.

That "Magic Inside" kind of thinking is a solid and much needed one. That's why the public is flooding toward the mobile appliances and nobody really bother to try the FOSS alternatives (Plasma, Meego) - no one is able to make sense of them because they're not task-oriented, at least not for the tasks that are more commonly needed.

But empowering users so that they can create their own magic, is an option not found at the new Post-Wimp style found at mobile devices. We need to find a way in which some of the possibilities of the Coherent Universe can be added to the task-based Magic Inside devices. That's the challenge for UxD and interaction designers in the next years.