 Reversible Garbage Collection (and more)
started 9/11/2003; 1:00:59 AM - last post 10/30/2003; 1:07:51 AM
|
|
David B. Wildgoose - Reversible Garbage Collection (and more) 
9/11/2003; 1:00:59 AM (reads: 482, responses: 12)
|
|
|
The Thermodynamics of Garbage Collection
A fascinating and readable account of some of the implications of Reversible Computation, including references to Psi-Lisp, a Reversible Linear Lisp.
Example snippet: The ability to summarize the information within a fixed number of "state variables" is the reversible computer's analogue to the "tail recursion optimization".
|
|
Patrick Logan - Re: Reversible Garbage Collection (and more)
9/11/2003; 5:40:17 AM (reads: 472, responses: 0)
|
|
|
Oh, we could spend a good bit of time and space on Henry Baker's papers.
It's been a long time since I read any of those. What a collection.
|
|
David B. Wildgoose - Re: Reversible Garbage Collection (and more)
9/11/2003; 6:57:43 AM (reads: 466, responses: 0)
|
|
|
Your comment has a sarcastic tone.
Care to enlighten us?
|
|
Patrick Logan - Re: Reversible Garbage Collection (and more)
9/11/2003; 1:30:38 PM (reads: 428, responses: 1)
|
|
|
Your comment has a sarcastic tone.
Completely unintended. I meant to say Henry Baker's papers are broad and deep, and I am in awe of his command of the problems he thinks about.
I didn't have anything to say about this specific paper since it's been so long since I've read anything of Baker's. But I'd highly recommend people spend a great deal of time with his entire repertoire.
|
|
Ehud Lamm - Re: Reversible Garbage Collection (and more)
9/11/2003; 1:51:00 PM (reads: 433, responses: 0)
|
|
I agree. It's an interesting paper (I read it a long time ago, so I don't remember the specifics). Baker's papers are well worth reading (he even wrote about Ada  ).
I was sure Patrick wasn't being sarcastic, and I am glad to see I was right.
|
|
Bryn Keller - Re: Reversible Garbage Collection (and more)
9/11/2003; 4:51:23 PM (reads: 414, responses: 0)
|
|
|
Are there any existing, implemented, reversible languages available for experimentation?
|
|
David B. Wildgoose - Re: Reversible Garbage Collection (and more)
9/11/2003; 11:22:07 PM (reads: 407, responses: 0)
|
|
|
Sorry for misreading the "tone" of your comment Patrick. I had also posted a link on another list only to have a response that the paper was "bull" based on a reading of just the first few paragraphs. I wouldn't have minded so much if this "Slashdot" approach to commenting without first reading hadn't come from someone whose opinions I usually respect.
I have now heard of 3 reversible languages, Psi-Lisp, Janus (nice name that one) and "R", but I don't know any more details.
|
|
Michael Vanier - Re: Reversible Garbage Collection (and more)
9/12/2003; 12:53:30 AM (reads: 394, responses: 0)
|
|
|
I was struck by the property of Psi-Lisp that procedures consume their arguments. I immediately thought of Forth, where this is the normal convention. Perhaps Forth could be used to make an accessible reversible language? You'd have to get rid of DUP and DROP.
|
|
Patrick Logan - Re: Reversible Garbage Collection (and more)
9/12/2003; 5:50:31 AM (reads: 384, responses: 0)
|
|
|
I was struck by the property of Psi-Lisp that procedures consume their arguments. I immediately thought of Forth
In fact Baker (not sure about this paper, but others) uses Forth as a means of illustrating linear objects.
|
|
Michael Vanier - Re: Reversible Garbage Collection (and more)
9/12/2003; 7:25:43 PM (reads: 364, responses: 0)
|
|
|
Patrick, do you offhand remember which papers discuss Forth? I can search for it, but I'm lazy ;-)
|
|
Michael Vanier - Re: Reversible Garbage Collection (and more)
9/13/2003; 1:17:25 AM (reads: 351, responses: 0)
|
|
|
Never mind, here it is:
http://www.pipeline.com/~hbaker1/ForthStack.html
It's worth reading. It deals with using a Forth-like (or PostScript-like) stack machine to implement a linear dialect of lisp. Linearity is an easier concept to grok than reversibility, although they're related. As I understand it, a linear language is one in which copying and duplication of objects have to be done explicitly. A reversible language is one in which copying and duplication simply aren't done. It's really fascinating stuff.
|
|
Shin-Cheng Mu - Re: Reversible Garbage Collection (and more)
10/30/2003; 1:07:51 AM (reads: 238, responses: 0)
|
|
I have now heard of 3 reversible languages, Psi-Lisp, Janus (nice name that one) and "R", but I don't know any more details.
I happen to know of another one. Paolo Zuliani described in his DPhil thesis a language based on Dijkstra's guarded command lanugage. The relevant chaper was published as a separate paper. It's more a language for specification than for implementation, however.
I am curious to know other related work.
(Sorry about the preivous message. I registered with a wrong email address..)
|
|
|
|