The article makes a lot of sense, and reflects practices that are well known, if not practiced all that much in the industry. I have worked for companies that use contextual design as described in the reading and can attest to the fact that it works well.
The ease in which contextual design practices can be implemented depend on (1) how far away from the current practice the final product winds up being, and (2) how well the users agree on what is being done and how it should be done.
I bring up the latter point because I am currently working on a system that started out as excel spreadsheets and has since been turned into a webapp. Initially, the users couldn’t see past making the system look just like a spreadsheet – in particular, their spreadsheet, and each user had a way of handing the information that was different. Much of the effort in the design and implementation of the system was in getting the users to agree on best practices. Interestingly, the further the system got away from the spreadsheet model, the more the users were able to look at the problem dispassionately and work together to arrive at a “standard” best practice.
With respect to the first point, I wonder a lot about self-driving cars. Building the technology that can actually get a car safely down the road is a pure engineering challenge. Cars do not drive the way that people do, and looking to how people drive is of limited utility in helping the development of this technology.
But users will sit in these cars. So how does that interface get developed? Do we look to chauffeurs? Passengers? Where’s the context that makes sense. Or is this the same as asking a society based on horse-drawn vehicles to describe how to interact with airplanes?