First, a general comment: The subject of the article was Fitts's Law, not cognitive processing time, not logical grouping of elements, nor any other of the many important elements that go into design. "Point by point:" 1. People don't read words like Save or Load after a while. Instead they are seen as symbols. Whether people can process symbols faster than icons is something I don't know. However, we had a saying in the Mac group a year or so after the Mac first came out, based on people's inability to deal with dozens of icons: A word is worth a thousand pictures. 2. See my general comment. 3. "A one pixel target? What is he smoking anyway?" I routinely have to hit one-pixel-wide targets. They lie, for example, on either side of the letter "i," and are amazingly difficult to hit with accuracy, so much so that I've taken to clicking in the general area, then zeroing in with the cursor keys when editing. This is a clear sign of a Fitts's problem, since deciding whether the user needs to select the forward or back arrow to solve the inaccuracy typically takes more than two seconds. 4.Logic is not the goal; ease of learning and productivity are the goals. Programmers often fall into the trap of making an interface that is perfectly logical, but all but unusable. An example occurred on the original Xerox Star, back in the early 1980s. Dragging a document from one device, such as a disk, to another moved the document. That it, the document was cut from the original device and pasted to the new device. This was logically consistent with the behavior within a device, where you don't leave a copy of the document in one folder when dragging it to another. It is, after all, illogical to sometimes copy and sometimes move a document when the user has performed the identical action, particularly since the user may not be aware they are crossing a device boundary. This system worked most of the time when crossing disk boundaries. True, sometimes people ended up taking a document home without appreciating the fact that they no longer had a copy at work, but they had been, after all forewarned. The real problem came when they moved the document from a disk device to a printing device. In the original Star interface, when you did that, you moved the document to the printer, and no electronic form of the document remained. Why? Because that was the logical thing to have happen. Fortunately, a more important principle--consistency with user expectations--soon won the day. 5. "Sensible" is right up there with "logical" as a slippery slope to be avoided. Users switch modes as they go from application to application. The menu bar switches with them. They never experience the sense of a menu bar that is "always switching around," because it is always in the same mode they are. The point about people wanting what they are used to is a good one. People, particularly post-school-age people, tend to avoid learning new things. They will put up with something less than optimum if it means not having to change their habits. 7. Since the words soon are scanned as symbols, and since motor memory soon learns the direction of travel, circular menus have positive benefit in habitual-use applications. It is not necessary to bend the words around like a pretzel, either. The targets are often big enough that the words can be written horizontally within them. If not, there are standard graphic techniques for handling letters on a curve in such a way that the words are readable without standing on one's head. 8. The wonderful thing about Fitts's Law is that it readily responds to user-testing. It is not enought to not "think" it is a good thing. Test it! 10. "Overall, I agree with some of his conclusions. Bigger buttons are probably a good idea. But to think that it all comes down to some kind of "mousing olympics", that it's all sprints and hurdles instead of a mental process, is plain dumb." I couldn't agree more. Of course, I never argued that it was. Fitts's Law is analogous to providing a car with accurate steering. Of course, driving a car entails a whole bunch of things as important or even more important that the mechanics of turning the wheel. On the other hand, if the steering is not fast, responsive, and accurate, you will be in trouble, even if your mental processes are all intact. -bruce tognazzini