Hello altogether,
introducing labeld fields into Curry seems to be a bigger topic than I
would have expected. The given papers, especially Daan Leijen's, show
that one could have a much more general approach without extending the
language in any other way (no need to introduce type classes for
example). On the other hand it is not entirely clear to me why the
Haskell98 approach effectively hindered the development of better solutions.
However, judging from the recent discussion I would propose that we
should not officially introduce record-like structures to Curry until we
have got some experience with the alternatives. We will therefore
implement a (not necessarily strict) subset of Leijen's proposal into
PAKCS and you can then all see how well it works in practice. After some
evaluation we could then again discuss what to take into the Curry report.
What do you think about that?
Greetings
Bernd
_______________________________________________
curry mailing list
curry_at_lists.RWTH-Aachen.DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/curry
Received on Mi Mär 01 2006 - 16:41:46 CET