Daisy is a statically-scoped, demand-driven, applicative
The design of Daisy is based on suspending construction,
a model for demand-driven parallel computation developed at Indiana
In the geneology of languages,
Daisy is a descendant of pure Lisp and is a contemporary of
In syntax and semantics, however, Daisy differs significantly
from most Lisp-family languages.
What's so great about Daisy?
The research goals behind Daisy can be summed up as follows:
- The language should automate concurrency and I/O management with
the same ease that modern languages handle memory management.
Just as garbage-collected languages simplify programming by removing
allocation and deallocation hassles, languages can (and should)
also automate thread creation, scheduling, synchronization, and
termination. They should also open and close devices
automatically as needed, buffer data efficiently and transfer data
directly and transparently into the heap.
[Eric Jeschke contact info]
Last modified: Sun Dec 21 14:14:22 HST 2003