In the Photoshop newsgroup, we sometimes discuss interfaces. After all, many of us there are designers; we work on computers, we see a lot of varied interface designs, and when we work on web pages, we are interface builders as well. On the web, there is so much freedom, to create your own paradigms, to create your own modes of interaction.

Photoshop has one kind of interface, built on established GUI standards, complex, tilted toward ease of use but definitely away from ease of learning. Kai's Power Tools has a completely different kind of interface. Each tool has its own, and they are often wild, imaginative designs, with controls not readily obvious, controls whose scales are invisible even when you are using them. The Spheroid Designer is one of the most provocative; it looks rather like a handful of marbles dropped in a mudpit.

People either love it or hate it. Those who love it fall in love with its graphical design. Those who hate it dislike its hidden features, its lack of repeatability, the fact that it hides your work from you, the fact that you have to think about the interface rather than your design.

I came into the fray from a human factors viewpoint. Yes, the current GUI standards can stifle innovation. But while Kai's Power Tools tries to break away, it's still mired in the same trap of windows, icons, menus, and pointers. The WIMP interface still rules.

To prod people into thinking, I suggested an immersive paradigm. What if you could create art by shaping a three-dimensional object in midair with your hands? What if the computer would show you your object even as your hands continued to shape it?

I got what I deserved. One person asked, "Why is waving your hands around in the air a paradigm shift?" Someone else answered, "For those expert in modeling clay it might be."

I have a close friend who is a potter. As she describes the pottery experience (she owes me lessons, I hope to collect on this someday), it is very much a tactile experience. She might say that the intelligence is in her fingers, and her fingers "see" by feeling the clay, its texture, pliability, responsiveness to pressure, shape, moisture, many things. She would not be happy waving her hands in the air, but then again a computer isn't going to make her pots, either. Part of what matters to her is the human touch with the clay, and it's fascinating to hear her discuss Japanese pottery styles, which often include carefully sculpted imperfections to ensure that the human touch of creation comes through to the person who eats or drinks from those pots.

Another close friend is an artist; she looks down her nose at computer art. Part of that has to do with computers, which bore and irritate her, but part of it also is that she, too, needs to feel the materials of her art. For her, creating art is a physical act.

She asked me once if I had ever created anything with my hands. I could truthfully say that I had; I used to ride bicycles, and I built the wheels that I rode on. I was a good wheelsmith. But I knew immediately what she was saying. She makes art; she makes bread; she makes clothes. She makes things with her hands, and that connection through touch is vital to her. As a software designer and a writer, I make things with my thoughts. There may be a physical result to what I make, but most of the time it is only patterns of light and darkness that still require a mind to interpret.

Obviously there are a lot of us who are quite comfortable creating art with a computer. At this point, virtually all of our feedback is visual, except for those who use a tablet in place of a mouse. Even the mouse could be enhanced; several years ago I had a chance to try out an experimental mouse with tactile feedback. There was extra drag over dark areas, and you felt a bump with your fingertip when you went over a line. It was subtle enough not to be interfering, and actually quite an interesting experience. But we all do manage to get along without that kind of feedback.

So who would use these wave-your-hands-in-the-air interfaces? Some of us would. Some of us wouldn't. Some of us might, depending on the task. And the task does matter; if the output is physical, then the process may need to remain both physical and tactile.

It is more than just the task. We do task analysis, so that we can understand who does the work, and how they do it. But as we move on to new interface paradigms, do we also need to know how they perceive their work at a sensory level, how they connect with their work at a human, biological, living-animal level?

We need to know who uses our interfaces. As the interfaces become more pervasive and more perceptive, perhaps we need to know more about the people as well as about the technology, so that we can truly build interfaces that are as satisfying to work with as working with the natural world.