Arc: an opportunity
Some of you will have heard of the language Paul Graham's been working on, Arc. “Another Lisp variant?” you may ask, and you'd be right. However, there is one important thing which I feel goes beyond the language itself.Lisps are much of a muchness; other than being a Lisp-1 or Lisp-2, everything else is just a choice of words*. Common Lisp was defined before the prevalence of networks, XML, Internet-based code libraries (like CPAN and RubyGems), and other things modern programmers take for granted. Indeed, the lack of extensive libraries is one of Scheme's perceived failings, and sometimes quoted as a reason to “put up with” Python. However, I think that Python has another great strength, and that's a reference implementation. There is only one Python, and you know Python code will run everywhere, because every other Python (e.g. Iron Python) has a validator: the original.
Arc is an opportunity to get everything right in Lisp: to sort out, for the moment at least:
- Window system bindings
- A full standard library (including sort, Scheme!)
- A central repository, system definition language, and infrastructure for packages
- Networking, synchronisation, parallelisation, and other construct definitions
- A canonical VM or compiler
Think what strides would be made if every Lisper and Schemer who was dissatisfied with their languages worked together to build a truly modern Lisp.
It'll never happen.
And why not use Python? It's evolved, so it has syntactical and library cruft, and it just ain't a Lisp. No macros, for a start.
* A gross over-simplification, but please forgive me!
Posted at 2004-10-22 10:44:26 by Richard • Link to Arc: an opportunit…
Comments, trackbacks.
