archives

Generic views

Stefan Holdermans. Master's thesis, Utrecht University, 2005.
An extension of Wadler's views.

In the conclusion, in a comparison to SYB, this paper notes:

Even when they target the same types of applications, it is often hard to transfer ideas and techniques between two different [generic programming] approaches. Therefore, frameworks for generic programming are typically studied in rather isolated environments. When proposals are compared to one another, the focus usually is on issues like which programs can and which programs cannot be written within the context of a specific paradigm: attention is drawn to the surface area of a generic-programming model, rather than to the underlying theory and techniques.

Generic views on data types may allow us to break with this tradition and incorporate different approaches within a single framework. The idea is to define a view that captures the way in which a given approach treates data types and to consequently transfer the building blocks of that approach to Generic Haskell idioms.

Before noting the impossibility of even this approach to encode generic fold.