On 1/3/08, Julian Rohrhuber <rohrhuber_at_uni-hamburg.de> wrote:
> At 17:01 Uhr -0800 01.01.2008, James McCartney wrote:
> >Conal Elliot came to Apple several weeks ago and gave a talk about
> >"tangible values". It was interesting. I think there is a complexity
> >limit for that approach.
>
> from my experience, I agree, but I am not sure. Why do you think this is so?
First, it forces you to think about programming as a process of
binding a single argument at a time. I often think about expressions
as an ensemble of function calls, not one argument binding at a time.
Second, as one of the audience brought up at the end of the
presentation, there is no means of introspection of the history of the
composition and he only had a guess as to what a solution might look
like (replay a "movie" of the composition).
Kyma originally ran into the same two issues above and solved it by
implementing the graph view of objects, i.e. back to boxes and wires.
Kyma has a longer history and a track record of usage by a user
community driving features, which an experimental system like this one
lacks.
Third, I don't see how free variables are expressed in this paradigm,
which is an important feature of lambas in functional programming.
Perhaps I am just missing something.
http://en.wikipedia.org/wiki/Free_variables_and_bound_variables
However even given some limitations, having end user composable UI
functionality would be a big step forward.
--
--- james mccartney
Received on Thu Jan 03 2008 - 18:13:49 GMT