Function types syntax

[Neal Gafter]

On Wed, Jan 27, 2010 at 6:42 PM, Mark Mahieu <markmahieu at googlemail.com> wrote:
> I know Neal warns against designing language features based on specific use cases,
> but it is interesting to note the connection between a particular perspective (or
> emphasised category of use cases) and choice of syntax, transparency or not,
> conversion to interfaces/abstract classes etc.

Analysis of language features by reference to use cases is very
helpful.  In this case, it is quite appropriate.  This analysis helps
ensure that the feature is composible with itself (higher-order
functions) and with other features (arrays).  That's not the kind of
thing I warn against.  I warn against defining the language feature
(or its interaction with other language features) by reference to the
behavior you'd like in particular use cases.  The complexity of a
language specification grows linearly with the number of features if
the features are orthogonal.  It grows as a higher-order function of
the number of features if the interactions of features are defined on
a case-by-case basis.