Giuliano Procida writes:
| Working from an unpublished paper by Sean Bechhofer we have developed
| a system of extensible records for Gofer/Haskell. This is based not on
| subtyping but instead uses the type class mechanism to handle
| constraints on the existence of fields.
This subject comes up with surprising frequency!
I developed a system of extensible records based on class constraints
rather than subtyping as part of my work on qualified types. This was
described in a couple of places, including this mailing list, and my
thesis (1992) which has now been published by CUP. My work was in turn
inspired by an explicitly typed system (dating from around 1989-1990, I
believe) that was studied by Harper and Pierce. Independently, Ohori
published similar ideas (but without making the connection with type
classes) at POPL 92, and has distributed an implementation (a modified
version of SML/NJ).
For further information on my own work in this area, browse the
publications section of my home page:
http://www.cs.nott.ac.uk:80/Department/Staff/mpj/
You will also find references to the other systems that I have
mentioned in some of the papers there.
| We have an implementation of a parser and type-checker, including all
| Gofer's features (even constructor classes).
The system I proposed predates the introduction of constructor
classes. To extend my proposals to work with constructor classes, you
need to adopt a polymorphic kind system with subkinding constraints.
Perhaps this is what you have done? I would be interested to hear more
about your work.
| Finally, when and where will FPCA be?
FPCA will be held in La Jolla, CA in June. The deadline for submissions
has now past. Further information is available from:
ftp://ftp.dcs.glasgow.ac.uk/pub/fpca95/fpca95.html
All the best,
Mark
