Data analysis is not a crime Web sites (applications) will “act” and “adapt” to user interaction as the years (days) go by … Resistance is futile.

If we take a humanistic view of human nature, and see people as interpreting, sense making, autonomous individuals in constraining social settings, who seek to grow throughout their lives, then 'nudges' apart, we can't really expect to fully influence human behaviour through design. Furthermore, the respect for individual autonomy here does raise ethical issues about covert manipulative behaviour. As long as the 'nudges' in an interaction design are overt and require some degree of knowing co-operation from users, then we can reduce the ethical issues.

Mark Steen (TNO, Netherlands) examined self-other constructions in his Design PhD. All design activities create a relationship between the designer(s)' self and the users' other, and also co-construct both self and other in the process. I find Mark's analyses very helpful here in reframing the sort of questions above. For me, the question is, what are the ethical relations between designers and users in terms of self-other relationships, especialy when the relationship is mediated by a design artefact?

Ultimately, it comes down to how explict designers are in expressing their purpose for designs, that is the ends to which a design is a means. Such purpose will vary, and have different ethical implications in different contexts, e.g., ATM dialogue design intentionally constrains the user to minimise queuing times, but a time-out on a print dialog would be seen as unethical – there is no need to pace users here, it's their PC and they an leave a dialog hanging as long as they want (this is not the same as Adobe/browsers *hiding* the print dialog under a pile of windows and freezing up your internet access!)

This may be a good process for successfully phasing out legacy features.

The structuring approach would provide users with a predefined “script” or “routine” helping them to go through the actions they have to accomplish towards a specific goal. This works well with new-comers or beginners as a “scaffold” for their activity. (in the same manner than adding additional wheels to a child's bicycle when he/she is learning to ride). The drawback is that it usually prevents users to be creative or to transfer solutions that already work form them.

The regulating approach is to provide users with freedom about their processes of interaction with the interface. In addition, one may add some sort of feedback support to help them figure out ways to efficiently reach there goal. This approach is less constraining and leave room for idiosyncratic routines as well as creativity in terms of optimizing these routines. These approaches would fit best intermediate and frequent users.

To sum up, I agree with your position but would add that it is dangerous to adhere to one of these school of thoughts. Both could have there advantages and drawbacks. Carefully analyzing the context of use and the specificities of the audience is the key.