A Jef Raskin Interview From A Year Ago

I only just found out that Jef Raskin passed away last month. I thought I would post an email interview I had with him a year ago to illustrate some of his thinking in his last year:

On Mar 9, 2004, at 7:22 AM, Jeremy Wagstaff wrote:

Jef, sounds better if I send the questions by email… I have greatly enjoyed your book, a real eye-opener, although unfortunately time constraints may mean I am not able to digest as thoroughly as I should have. So please forgive any questions below which could be answered by a closer reading of your book! — Would you mind giving a brief elaboration of your comment that ‘I have learned a great deal about interface design, human psychology, and human physiology since creating the Macintosh project a quarter century ago — and even then I wanted to use the mouse far less than the larger role given to it by later workers at Apple’. What have you learned, exactly? Where are we going wrong with the use of the mouse?

To ask what I have learned, exactly, is answered by my book and articles. There is no short answer that fulfills the requested “exactly”. It is now well established (I sent you the latest and best reference) that mouse use should be minimized, and it has been long known (since at least the 1980s) how slow mouse operations are. The problem has been that keyboard-based solutions have been even worse in terms of learnability and memorability. But the mouse (or other pointing device; I prefer tablets for drawing and a good trackball for pointing, but that’s personal preference) is essential for graphics. THE is designed to use both the keyboard and the mouse where they are appropriate and not use them where they are not; and I have found pleasant solutions that make THE both learnable and memorable — as testing has shown.

But the most important things I have learned are those involving how humans learn and work. Applying that research-based knowledge and using the quantitative tools that have been developed facilitate the development of much better interfaces than we now have. – When you say ‘It is still the case that most of what we do with computers (estimates are typically 80% to 85% on a time basis) involves the creation, reading, and editing of text. And for this kind of work, the mouse is usually inappropriate’ could you give examples? Do you mean users should use more keystroke combinations?

Using the rather arbitrary keystroke combinations now available is a finger-twisting exercise that is frustrated by the inconsistent way they are used in different applications. Often there are no keyboard methods for some tasks. The present ad hoc keystroke combinations are pathetic. Users should demand better software, there is no way to use the present interfaces well.

I was going to suggest in my column that users make a better distinction between tasks, i.e. between a) thought flow tasks — writing, mainly — where reaching for the mouse could only distract and disrupt thought, and therefore should be avoided and b) housekeeping tasks, where the visual GUI could be made more use of via the mouse than it is — for example, by dragging things between windows. But perhaps I’m barking up the wrong tree, or confusing people further?

I don’t think that that advice makes sense. For one thing, most housekeeping tasks should also be keyboard-based.

 As an ‘occasional’ RSI sufferer I’m fully with you on the mouse issue, and have recommended on a couple of occasions to cut back on mouse usage. So I worry that with the above recommendation I might be making things worse. But while there are some great programs out there which can significantly cut down on mouse usage to do ‘housekeeping tasks’  — I can’t see any way of avoiding some tasks altogether. Or am I not thinking far ahead enough?

Another question that requires a longer answer than you can possibly use. Sometimes difficult problems don’t have answers that can be put into a pithy paragraph or two. The specs for just the text portions of THE run to over 40 pages, and in those pages are very specific answers to your questions. I don’t know about “ahead” but you are only beginning to think outside the restrictive GUI box. It’s hard to understand a new world when you’ve spent years in the old.

 Finally, leading on from that, what will the interface of the near future look like? What can people expect, and how can they help make that day arrive sooner? Are there any specific tips you could offer users who don’t want to wait for changes in their GUI?

In the near future, people will use today’s GUIs. Taking a longer term view, I hope that developers will read books, of which mine is but one example, about how people really interact with computers and other similar products instead of using the present half-facts or outright false beliefs about what makes an interface work. If I find the support and/or sufficient volunteers, I will get THE out into the world, and people will gradually move to it because it is significantly better.

I can’t help current GUI users; I search through my Windows for Dummies and Macs for Dummies type books and try to learn how to use them effectively, but they are so wrong from the getgo that there’s only so much a user can do to make them better.

A Short Essay From Jef Raskin

Further to the previous post, honouring the fact that Jef Raskin passed away last month, I thought I would post a little essay he sent me a year ago to illustrate some of his thinking in his last year:

Genesis and Goals of The Humane Environment

Our increasing knowledge about human behavior and mental processes, as applied to interaction with our artifacts — knowledge based on observation, on testing, and on empirical results in cognitive psychology — leads to the conclusion that the human/machine interfaces of current computers, cell phones, PDAs, automobiles, and much more are often flawed. Their interfaces features often derive from faulty precedents, and on inadequate models of and incorrect informal guesses about human performance. In particular, GUIs (such as Windows) used by hundreds of millions of people reflect these problems in abundance.

A more accurate external model of human mental processing leads to quite different interfaces than those we now have. One approach to applying this knowledge has resulted in “The Humane Environment” project. There is no reason to believe that it is the only approach or the optimal one, I do claim that it is considerably better than current practice or alternatives of which I am aware in terms of speed of learning, productivity, and the feeling of trustworthiness. Not only applications, but programming languages and software development systems are also human-machine interfaces and their design can benefit from developments in cognetics. (Cognetics is the engineering of products to accommodate human mental abilities and limitations; an analog of the better known ergonomics, which guides the design of products to match human physical attributes.)

My background has biased me toward that which is quantitative, deductive, empirical, practical, and humanitarian. Applying these criteria reveals that only a small fraction of books and articles on interface design are applicable to development in any rigorous sense; most are hortatory, few get beyond offering heuristics, many are irrelevant or simply wrong. The quantitative tools that are available in this field are unknown to a majority of practitioners, as I discover nearly every time I give a lecture to audiences of professional or academic HCI practitioners — a situation that I find deplorable. (My evidence comes from asking people at my talks whether they know this or that quantitative method. Usually only a few hands are raised). The HCI research literature is often pathetic, with poor experimental design and overblown conclusions. Very common are studies that compare a particular instance of technique A that is superior to an instance of technique B. They then conclude that technique A is superior to B; ignoring that it may have been a great example of A and a very poorly implemented B: Conclusions that go beyond the premises is a common error in the field.

The weak research and the widespread belief that the way computers are is how computers must be, coupled with the bias toward standard the GUIs built into current operating systems and development environments, has stymied progress. The importance of habituation and of our single locus of attention, for example, have not been widely recognized.

The theoretical reasons for believing that THE is an improvement over current designs are very strong, and equally strong is our experience with the SwyftWare and Canon Cat products that embodied the principles and some of the technology of the text portion of THE (which is inherently usable by the blind). The zooming interface implemented at Apricus Inc. showed the effectiveness of the graphical portion. When theory and user testing meet in this way, and a refactoring of how computers should be used yields a much more compact design while offering users and programmers greater power than present systems, I have considerable confidence in the work. Many people are also intimidated by their fear that any change from the Microsoft Windows way is doomed to failure because of its large installed base. Perhaps they have never heard of Linux, they are not entrepreneurial, they are doomed to nebishhood. Sufficiently better products can penetrate the marketplace.

Considering the millions of person-hours that can be saved, the mental toll of frustration that can be eased, and the physical pain that can be prevented by putting THE into the world, I feel compelled to work on and promote it — and to try to motivate those who can help to do so.