3 ms·

I'd like to see - maybe one year I'll have both the time and expertise to put work towards it - a Common Lisp that can be retargeted to different vms in a strai

by pnathan 16y ago

I'd like to see - maybe one year I'll have both the time and expertise to put work towards it - a Common Lisp that can be retargeted to different vms in a straightforward fashion: JVM, .NET, Rubinious, llvm, etc. While it's awesome to make yet another Lisp, I think it's far more awesome to have a well-defined Lisp that can run on multiple places.

bitsai 16y ago
That's one of the goals of the Clojure team. While the main 'branch' targets the JVM, there is also a branch targeting .NET. And there is an initiative, Clojure-in-Clojure, to de-couple the main branch from the JVM as much as possible, to facilitate re-targeting.
rbxbx 16y ago
Clojure's primary target is the JVM and they had an officially supported CLR(.NET) branch for a while (and it's now supported by a 3rd party https://github.com/richhickey/clojure-clr https://github.com/richhickey/clojure-clr). I wouldn't be surprised if in the future portability became a bigger concern within the Clojure community (http://blog.n01se.net/?p=41 http://blog.n01se.net/?p=41). (That said I realize you were looking for a Common Lisp, but sometimes good enough is good enough ;) )
cabalamat 16y ago
> a Common Lisp that can be retargeted to different vms in a straightforward fashion: JVM, .NET, Rubinious, llvm, etc. [...] awesome to have a well-defined Lisp that can run on multiple places But it wouldn't. A Lisp that ran on the JVM would be able to use JVM libraries, one that ran on .NET would be able to use .NET libraries, one that ran on Rubinious would be able to use Ruby libraries. Sure, the core Lisp language might be the same, but as soon as you started doing real work connecting it to outside code/services/devices etc, it'd rapidly become non-portable. That doesn't mean it would be useless: being able to think in one language over multiple platforms would reduce cognitive load for developers.
- pnathan 16y ago
  well, yes. But what would be able to happen in A Perfect World is for the intersection of features to have an abstraction layer.
papaf 16y ago
Bigloo Scheme is pretty cool and can target native code, JVM, mono and Android: http://www-sop.inria.fr/indes/fp/Bigloo/ http://www-sop.inria.fr/indes/fp/Bigloo/

by pnathan 16y ago

bitsai 16y ago

That's one of the goals of the Clojure team. While the main 'branch' targets the JVM, there is also a branch targeting .NET. And there is an initiative, Clojure-in-Clojure, to de-couple the main branch from the JVM as much as possible, to facilitate re-targeting.

rbxbx 16y ago

Clojure's primary target is the JVM and they had an officially supported CLR(.NET) branch for a while (and it's now supported by a 3rd party https://github.com/richhickey/clojure-clr https://github.com/richhickey/clojure-clr). I wouldn't be surprised if in the future portability became a bigger concern within the Clojure community (http://blog.n01se.net/?p=41 http://blog.n01se.net/?p=41). (That said I realize you were looking for a Common Lisp, but sometimes good enough is good enough ;) )

cabalamat 16y ago

> a Common Lisp that can be retargeted to different vms in a straightforward fashion: JVM, .NET, Rubinious, llvm, etc. [...] awesome to have a well-defined Lisp that can run on multiple places But it wouldn't. A Lisp that ran on the JVM would be able to use JVM libraries, one that ran on .NET would be able to use .NET libraries, one that ran on Rubinious would be able to use Ruby libraries. Sure, the core Lisp language might be the same, but as soon as you started doing real work connecting it to outside code/services/devices etc, it'd rapidly become non-portable. That doesn't mean it would be useless: being able to think in one language over multiple platforms would reduce cognitive load for developers.

pnathan 16y ago
well, yes. But what would be able to happen in A Perfect World is for the intersection of features to have an abstraction layer.

papaf 16y ago

Bigloo Scheme is pretty cool and can target native code, JVM, mono and Android: http://www-sop.inria.fr/indes/fp/Bigloo/ http://www-sop.inria.fr/indes/fp/Bigloo/