User Interface Design for Programmers: A Book Review
The last time I checked, software was meant to be used.
In order for software to be used, it needs to be useful to those who have not perused all the texts required for a Ph.D. in Computer Science--as well as to those who have. This is rather less common than one might wish in the Open Source community. Often, the applications written for Linux seem to be "by geeks, for geeks". The fact that we are many of doesn't mitigate the preference of Janey-or-Johnny-User for a consistent interface among the applications they use, and for those applications to behave in a way that is comprehensible to someone unfit for MENSA(R) membership.
User Interface Design for Programmers (Apress, 2001) is written for programmers, especially programmers who a) think user interface design is an artistic pursuit, b) have a sense of humour, and c) can no longer think like Janey-or-John-User.
Who? Us? Yes.
The reader of Mr. Spolsky's work will require both a sense of humour, and some acquaintance with the cultural context in which he writes. Lacking either or both, the reader will not go all verklemmt, but will go straight to meshuggah, neither passing GO nor collecting $200. This is not a book for those who think tech and storytelling don't mix. For those of us who like to smirk and learn at the same time, it's a wonderful read.
There are not a great many software engineers who would deem their work as being of signifigant importance to, say, mental health. Or human rights. Yet Spolsky begins with the premise that there is some relationship between being able to control one's environment and mental well-being, and the relationship between not being able to control one's environment and depression. Specifically, he relates how this applies to the users of computer programs:
UI is important because it affects the feelings, the emotions, and the mood of your users. If the UI is wrong, and the user feels like they can't control your software, they literally won't be happy and they'll blame it on your software. If the UI is smart, and things work the way the user expected them to work, they will be cheerful as they manage to accomplish small goals. Hey! I ripped a CD! It just worked! Nice software!" (p. 6)
The moral of this story? "A user interface is well designed when the program behaves exactly how the user thought it would". (ibid., emphasis mine) What? Not how we think it should?
Spolsky uses 18 short chapters to illustrate the various principles and pitfalls of UI. (There are some pratfalls, too, some of which involve UI designers from the Redmond Contingent, as well as from Juno). The author's anecdotes are lively, and generally point to a specific corollary or rationale involved in designing applications that "make users happy". There are worse goals in the development of software.
"Good UI designers use consistency intelligently, and though it may not show off their creativity as well, in the long run it makes users happier." (p. 48)
Spolsky might have added that this can be taken to extremes, as in the case of StarOffice's 'butterfly start button'. On the whole, though, he makes a good point. For those of us who would like to increase the number of Linux users, consistent interfaces make the job of helping newbies view Linux as more user-friendly. This does not preclude efficiency or security, by the way: the sooner the newbie can accomplish what s/he intended to, the less likely s/he is to look for an alternate command that will permit him or her to accomplish something unintended.
"Design for extremes so that your product can be used under extreme conditions, and design for extremes so that your product is more comfortable to use under normal conditions" (p. 59)
The good people who developed wheelchair ramps know this principle well. Slopes are easier to navigate than steps, for both wheelchairs and normal pedestrian traffic. Well-cushioned seats in vehicles are more comfortable and, in a bizarre twist of fate, safer than their harder counterparts. Food that doesn't take as long to cook retains more of its original nutrients...well, okay, maybe that's pushing it. Sometimes pandering to creature comfort and/or sheer laziness can be put to good use.
"A good heuristic is obvious, easily undone, and extremely likely to be correct. Other heuristics are annoying." (p. 113)
Microsoft Bob died. Many users of MS Office would prefer the spellcheck to simply let them make spelling mistakes and suffer the consequences. The Paperclip infuriates everyone over the age of 5. Don't create bells and whistles that require removal by less than obvious means.
"Let's do the time warp again"
One of the most salient points Spolsky raises is what happens to time between development and software use. Days of development yield seconds of software to the user; months spent on a project yield knowledge the user won't have in minutes of its use; seconds of time required to load an application (or a web page) feel like hours of boredom to a user. Lest we forget.
There are a few.
Mr. Spolsky assumes that the reader is a Windows programmer. This is a ridiculous assumption on several counts, none of which I will go into here.
(Let me rephrase that: there is a special place for those who would send a non-Linux book for review to a publication which deals more or less exclusively with Linux-related topics. It is a place where the walls are bright yellow, trees look like green lollipops on brown sticks, and there is no word longer than 5 letters. Authors of books don't, as a rule, spend longer there than it takes them to find something to mock.)
The illustrations in the book serve as a wondeful reminder of the reasons we "penguin-heads" have shunned software originating from The Redmond Contingent. Pretty colours, ugly interface, give us a CLI!
The Linux developer is pretty much forced to separate the principles of good UI design from the pictures, and extrapolate what they'd look like in WindowMaker, KDE, or even Enlightenment.
Spolsky states that he really doesn't think that all users are dolts, having spent 3 chapters convincing us that the majority of users are illiterate, (Chapter 9, People Can't Read), uncoordinated (Chapter 10, People Can't Control the Mouse), and are senile to a greater or lesser degree (Chapter 11, People Can't Remember). His "imaginary user" Patricia-the-Academic is an insane stereotype in a time when even non-technical people are learning to value things like security and stability. His attempt to present ease-of-use concerns with cautions against developer arrogance, while both are valid, merely sounds like rapid back-pedalling and detracts from what could have been very useful "words to the wise".
So what if Linux is perfunctorily mentioned as an OS with a certain degree of difficulty? It doesn't have to be. It's our job to change that.
User Interface Design for Programmers belongs in most software engineering departments. This is not an overstatement: the book presents some sane and grounded approaches to UI design which have nothing to do with aesthetics, and everything to do with usability. For hackers who have a good dose of creativity, the principles illlustrated in the book are transferable to any of the window/ desktop managers used on Linux. This, one can hope, will enable refinement of some potential "killer apps" that exist under Linux, and in turn increase the user-base of both the applications and Our Esteemed OS.
It is good you should know.
Author: Joel Spolsky