When I originally added Saffer to my reading list I suppose I was hoping for, among other things, an IxD definition of "interaction"; a view of what it was they were `designing', and perhaps trace amounts of a critical language for analysis of existing (and in progress) artifacts. This was not the place to look. Saffer provides explanations of common phenomena such as Hick's and Fitt's law and a brief section on interfaces. Most of the material is focused on process and issues of problem definition. Products (like the ipod) are presented primarily in the context of the process of their creation. Critical analysis of artifacts is slim. While impressing (rightly) that success can be highly dependent on the proper definition of the design problem(s) at hand , Designing For Interaction becomes effervescent when it concerns actual designing.
As a broad summation of interaction design it offers bits of pieces for everyone, but seems aimed at no one. There are breadcrumbs of potentially useful information for practitioners, and overviews of topics that are likely no interest to students (If i were confronting interaction for the first time, I'd be far less interested in how to align to buisnes strategies and far more interested in exploring foundational formal issues.)
What I found most interesting is the discussion it led me to ([...]). Early in the book while covering various design methodologies Saffer touches on the `Genius' approach to design. Contrary to (various forms of) User Centric Design which place a strong emphasis on research and pre-production work, `Genius' is described as a process that relies on the exceptional experience and skill of an an individual or small group of designers above the preliminary qualitative research that UCD seems to hold sacred. There's then a brief interview with James Leftwich where James proposes the term `Rapid Expert Design' as an alternative to avoid various connotations of the latter. In both the book and online he attempts to describe the differences between RED and UCD related practices. Unfortunately both distill to common design practices in more established design fields; building experience through apprenticship and projects of increasing scale then increasingly relying on said experience to make and explore design decisions intuitively and quickly. To Saffer's credit, his shorter distillation of the `Genius' approach is probably as descriptive, brief, and fair as can be.
Luckily the conversation is advanced by the likes of Jonas Löwgren who manages to hit not only the thread's nail on the head, but put words to what's been bothering me about interactive design for years (and perhaps lacking in Designing for Interaction).
"As I read Jim's discussion of RED, the key is the abilities [opposed to methodologies] that the RED designer holds... A general problem in developing design ability is the relative inefficiency of the learning process. Apprenticing and peripheral participation is the most common strategy and it generally takes a long time to reach expert levels of experience and performance... Does the RED approach contain any provisions for increasing the pace of learning? Do you work systematically with product reviews and criticism in your teams? Do you have procedures for debriefing and knowledge sharing after project milestones and completions? How are you working with conceptual tools for articulation of practical knowing, such as patterns or experiential qualities?... I was thinking also of language constructs for talking about what constitutes good interaction. The way I see it, this is one of the main elements of interaction design expertise (the "experience" we talked about earlier in this thread) and my personal approach is to try and articulate so-called experiential qualities to try and create a language in which experienced designers can express and communicate parts of their judgment skills." - Jonas Löwgren
As passionate as Leftwich is about RED, and as measured as he is in expressing his points. He unfortunately lacks the critical language to articulate his (and others) experience. There's nothing wrong with mentoring, but there's nothing good about each generation learning everything through trial and error. Knowledge needs to be codified so it can be, at minimum, passed down. While structured methodologies like UCD that focus on preliminary research are valuable, they contain little design knowledge in of themselves.
While I don't fault Dan Saffer for somewhat neglecting this issue, I imagine that moderately skilled interactive and interaction designers with some experience will a little disappointed.