The Humane Interface was a worthwhile read. I recommend it to those interested in UI design. However, I also recommend that you take the book with a large grain of salt. Raskin gives good background on HCI and cognition, but he also writes about UI design decision that are his own untested or semi-tested ideas as if they are on par with the well established ideas he mentions. He makes many good points, but I often disagree with his justifications. Following are some points that particularly bug me.

Modes: Raskin is rabidly, religiously against modes. Anyone who knows my editor of choice (vim) might guess that this is something I do not agree with. Raskin argues that modes are universally bad, but he then goes on to say that what a mode is varies from person to person. His definition of a moded interface is something like, "an interface is moded when the same action does different things in what the user perceives to be the same context". This is somewhat valid. The mantra of HCI is that what really matters is the user's perception. However, it introduces a chicken and egg sort of problem. Users define different contexts by how they want to use them but also by the modes that are given by the UI designer. Furthermore, Raskin ignores that the real world is moded. When driving, the users focus is on getting from point A to point B not on moving levers and hitting peddles. However, a vehicle behaves differently when you hit the gas peddle depending on whether its mode is neutral, reverse, or 1st gear.

Modes can be damaging, but they can also be useful. They increase what one can do with a limited set of commands. Modes are okay when disjoint (no command performs the same action in different modes) and clearly marked. When a user does switch modes, it should be because they are changing the locus of their attention. By this definition, unsurprisingly, the modes in vim are mostly okay. The two primary modes are almost completely distinct. The modes are not clearly marked. The mode change corresponds with a task change (editing structure verses entering text), but it can be easy to change modes accidentally.

Raskin also discusses noun-verb verses verb-noun interfaces. The former occurs when the user selects and object and then applies an action. The later occurs when the user selects an action and then selects an object to apply it to. He gives three justifications for why noun-verb is better:

1. The object is what the user is manipulating and is the locus of their attention. Therefore, making it the primary actor in the action is better. I agree with this point.

2. A noun-verb paradigm has only one attention switch from the noun to the verb, then you are done. The verb-noun paradigm has two attention switches, from the noun to the verb and then back to the noun to select it. This may be true in theory, but I would guess it is not true in practice. I would guess that, in practice, even with a noun-verb paradigm, the attention switches back to the same noun once the action is performed, either to verify that the action was correct or to perform further manipulations on that object.

3. To cancel a noun-verb action after a noun is selected, no action is needed. To cancel a verb-noun action after the action selected, a cancel button is needed. Therefore, noun-verb is less work. This is is wrong. Raskin is applying noun-verb thinking to a hypothetical verb-noun environment. In a true verb-noun environment, no canceling would be needed. Suppose a user selects and action and then decides that she does not want to perform that action. She does so simply by not choosing an object. Now, consider the next action she performs. Since, this is a verb-noun system, she will still have to choose the action before the object. Thus, she just selects a new action and then select the object to apply it to. Raskin's critique would be valid in a mixed verb-noun, noun-verb system. If a user chose an action and then decided not to apply it to an object and then decided to perform a noun-verb action, the user would have to cancel the action to be able to select the noun.

That's enough of noun-verb, verb-noun excitement. Raskin was often inconsistent. He thinks we should get rid of file systems because they impose a one-size-fits-all system on the user. He said, "one advantage of filing information as you wish is that the structures were not dictated by the systems designers, who may have ideas different than yours." However, Raskin claims that users should not be able to redesign, i.e., customize, interfaces. He never justifies why users are qualified to design file systems but not qualified to design interfaces.

Okay, that ranting felt good. Raskin's book was interesting but frustrating at times. He makes good points about simplicity and about considering the users locus of attention (hint: it's their task, not the interface). However, he tends to wander into the realm of fancy and provide inadequate justifications.

The Humane Interface del fu Jef Raskin è un classico testo sull'interfaccia uomo macchina.Raskin fu il padre dell'interfaccia del computer Apple Macintosh. Il punto di vista di Raskin era, ed è ancora, innovativo; le sue idee, se applicate, renderebbero l'uso del computer più piacevole e produttivo. Però questo libro non è di facile lettura: i capitoli sono troppo lunghi, l'inglese è spesso troppo letterario e certe volte manca una logica chiara su dove siano trattati certi argomenti. Alla fine della lettura si rimane amareggiati nel constatare quanto i nostri computer e le applicazioni che tutti i giorni usiamo siano ancora così lontani da quello che Raskin descrive. ( )

A disappointingly large amount of this book is devoted to Raskin selling the reader on features of his interface for the (failed) Canon Cat computer. There is some interesting material about how to evaluate interfaces, and some interesting ideas about task-focused computing (as opposed to os/application-focused computing). ( )

All computer interface designers should read this book! It's full of things that you see and say "of course, why didn't I think of that?" and plenty of others that make you think "why does everyone keep doing that so badly?" ( )