Cynthia Selfe and Richard Selfe, in "The Politics of the Interface," address the instrumentalism that prevails in critical discussions concerning information technology, by pointing to the implications (in Haraway's sense) of employing such technologies in the classroom. Their work serves as a reminder that the technologies that many critics treat more or less as neutral tools, do in fact instantiate powerful assumptions concerning the ethnicity of the user, as well as his or her gender and class.
Just as significantly, information interfaces create models of the appropriate uses for information, the correct processes to bring to bear upon that information, and the kinds of domains where such skills can be used. Selfe and Selfe point to the very restricted view of what constitutes information that is encoded in many interfaces, and their inflexibility in terms of customization for the needs of users. In fact, what Selfe and Selfe are doing in recognizing that interfaces are not value-free but are implicated in the power struggles and technological protocols of other domains, is to challenge the conception of the "user" that is inherent in the more instrumentalist rhetoric of technology critics.
The "user" is often seen as a discrete entity, easily differentiated from the technology that he or she is using. The user is also accorded a degree of expertise and prestige as a result of their knowledge. In fact the user posited by instrumentalist rhetoric is essentially a passive construction that masquerades as an active one: if there is expertise it is more often expertise in a particular interface or tool-using scenario; and "use" is more closely allied with a passive consumption of existing solutions than creative problem-solving or critical intervention.
Selfe and Selfe stress that for teachers to be effective agents in an increasingly complex, multi-faceted informational domain we need to cultivate flexibility as well as intuitive leaps based on holistic understanding of the structure and purposes of informational systems, rather than adhering slavishly to the linear thought processes encoded in particular interfaces. To this end we need to break away from such a rigid instrumentalism that conceives of informational subjects only as tool users rather than tool shapers. Hence, the authors argue, we need interfaces that are readily customizable to reflect not only the user's needs (language requirements, for example), but their previous experience in other informational domains and their preferred methods of problem-solving (495-6).