Talk:Actor model, mathematical logic, and physics

From Wikipedia, the free encyclopedia

Note: For a related, past dispute about the content of this article, see this talk page section. Karol 19:54, 25 October 2005 (UTC)

Update to above note. We are still working on the above issues in Incompleteness of quantum physics.--Carl Hewitt 21:06, 25 October 2005 (UTC)

Contents

[edit] Initial reorganization

The article has been reorganized as a result of the helpful questions, comments and suggestions that everyone has provided.

Note that current controversy seems to be confined to the section Actor model, mathematical_logic, and quantum physics#Actor model and quantum physics which is now completely independent of the rest of the article.

So one possible way out of our difficulties would be to delete the section Actor model, mathematical_logic, and quantum physics#Actor model and quantum physics and rename the article Actor model and mathematical logic.

If we did this, it would make sense to remove the connections between the Actor model and quantum physics from the other articles as well. Note that we would still have a connection between the Actor model and indeterminacy because the consensus view of experts is that Arbiters are indeterminate in their operation. Also it still leaves the Actor model dependent on physics but not yet reported that it is dependent on quantum physics.

Please provide questions, comments, suggestions, etc. Thanks--Carl Hewitt 10:18, 20 October 2005 (UTC)

[edit] Reorganization as per suggestions of Wikipedia editors

The article has now been reorganized as per suggestions of the Wikipedia editors. Note that it no longer mentions "quantum."

Therefore I suggest that the article be renamed Actor model, mathematical logic, and physics--Carl Hewitt 04:39, 21 October 2005 (UTC)

[edit] Reorganization of this article

Once the VfD is settled, if the article is not deleted, you will have to make substantial changes to this article. You may want to ask someone else such User:Chalst for assistance. Some suggested changes

  • The opening sentence has to be replaced. It is too vague.
  • "Hewitt's thesis was that logical deduction was incapable of carrying out concurrent computation in open systems even though there are mathematical theories of concurrency" This requires explanation. Surely you are not trying to claim that in principle the classes are different. You need to be more precise here. What is an "open system"; do you mean "open" in sense of its being entangled with the environment?

--CSTAR 17:09, 23 October 2005 (UTC)

Thanks for your comments. I have clarified the article. Please see what you think. Thanks,--Carl Hewitt 05:23, 26 October 2005 (UTC)

[edit] Grammar

The following text is not gramaticaly correct

For example Arbiters can be used in the implementation of the arrival ordering of an Actor turn give rise to physical indeterminacy.

Please fix it, so that I can see if you are trying to say what I think that you are trying to say. --EMS | Talk 15:41, 24 October 2005 (UTC)

Thanks for noticing this. I have improved the wording as follows:
For example Arbiters can be used in the implementation of the arrival ordering of an Actor which can give rise to physical indeterminacy in the arrival order.
Your comments, suggestions, and questions are greatly appreciated.--Carl Hewitt 05:30, 26 October 2005 (UTC)

[edit] Mathematical Logic

I dispute the claim made here that mathematical logic (or, to be precise, graph theory) cannot model the Actor model. This may be OR on my part, of it seems to me that a Kripke structure could be used to model the set of possible states of such a system, barring only the possibility of simultaneous Actors acting on the same data node. Arthur Rubin (talk) 00:55, 26 October 2005 (UTC)

Thanks for your comments. I have clarified the article. See what you think.--Carl Hewitt 05:11, 26 October 2005 (UTC)
Better. There's still no physics, though. Arthur Rubin (talk) 01:31, 28 October 2005 (UTC)
The article states:
The Actor model makes use of arbitration for determining which message is next in the arrival ordering of an Actor that is sent multiple messages concurrently. For example Arbiters can be used in the implementation of the arrival ordering of an Actor which can give rise to physical indeterminacy in the arrival order. Therefore mathematical logic can not implement concurrent computation in open systems because of the impossibility of deducing arrival orderings since they are indeterminate. Note that although mathematical logic cannot implement general concurrency it can implement some special cases of current computation, e.g., sequential computation and some kinds of parallel computation including the lambda calculus.
In concrete terms for Actor systems, typically we cannot observe the details by which the arrival order of messages for an Actor is determined. Attempting to do so affects the results and can even push the indeterminacy elsewhere. e.g., see metastability in electronics and arbiters. Instead of observing the internals of arbitration processes of Actor computations, we await outcomes. Physical indeterminacy in arbiters produces indeterminacy in Actors. The reason that we await outcomes is that we have no alternative because of indeterminacy.
Regards,--Carl Hewitt 02:17, 28 October 2005 (UTC)
Your use of physical indeterminacy is at least not NPOV, if not OR. I would be perfectly happy with the article if you remove all physical references, even if I think the relationship to mathematical logic is incorrect. Arthur Rubin (talk) 17:26, 29 October 2005 (UTC)
I sort of hate to be on Carl's side of this, but I see little trouble with his referencing physical indeterminacy in this. Certainly the Arbiters which he is so fond of display it. My gripe comes when Carl makes the illogical leap from the actor model referencing physics to the actor model being physics. If he would not do that, there would be no fight over what is fundamentally a very benign relationship. --EMS | Talk 04:55, 31 October 2005 (UTC)
I question the relevance of Arbiters to the Actor model, as well. Arthur Rubin (talk) 07:07, 31 October 2005 (UTC)
Arbiters are mentioned on p 36 of Gul Agha's thesis. I don't know if they are ever mentioned again.
However, some of the claims about the physical basis of the actor model do not appear in either Gul Agha's thesis nor in Clinger's thesis. For instance, the following claim made in Actor model theory is as far as I can tell, not in any of the published literature on the actor model (which is not to say that Hewitt may not have made it in an unpublished seminar talk)
Because of the transmission of energy, the activation ordering is relativistically invariant; that is, for all events e1.e2, if e1 --act--> e2, then e1 precedes e2 in the relativistic frames of reference of all observers.
Contrast this with what is on p 27 of Will Clinger's thesis:
"The theory of relativity allows each observer his or her own global time. These global times may differ, however, concerning the order of events whose relation in time is not absolute.
There is an analogy with global time in the actor model..'.'".
Please read the full quote. Taking this analogy in Clinger's thesis (in which Clinger clearly says it is just that, an analogy) and transforming it into a consequence of a principle of physics (with energy thrown in) is questionable. --CSTAR 16:41, 31 October 2005 (UTC)
But then, I've fought against the claim made by structured programming evangelists that one cannot prove correctness of unstructured programs. Arthur Rubin (talk) 00:58, 26 October 2005 (UTC)
It's worth observing that there is a well-established argument in modal logic that the frame conditions corresponding to formulae of modal logic are not first-order definable, hence if the domain of mathematical logic is taken to be that of multi-sorted theories in a first-order logic, then there can be a respectable case for graph-theory based theories not be be within mathematical logic. One must be very careful with such claims, they are prone to be misunderstood. --- Charles Stewart 13:37, 26 October 2005 (UTC)
So it's (approximately) second-order mathematical logic, rather than first-order mathematical logic. It's still mathematical logic. Arthur Rubin (talk) 01:31, 28 October 2005 (UTC)

[edit] Edit war, anyone?

Everyone: Please see this comparison of versions for this article. I see my wording as being more correct. However, I am also certain that Carl will revert it. So I need help for an edit war. Without support, I let Carl win (since this is more his baliwick than mine). --EMS | Talk 04:38, 28 October 2005 (UTC)

[edit] Articles for Deletion debate

This article survived an Articles for Deletion debate. The discussion can be found here. -Splashtalk 22:30, 28 October 2005 (UTC)

[edit] Status of block paragraph

The article contains the paragraph which begins:

The Actor model makes use of arbitration for determining which message is next in the arrival ordering of an Actor that is sent multiple messages concurrently.

Is this a quote? If so from which of the two articles? If it is not a quote, then why does it have this special formatting? --CSTAR 04:45, 3 November 2005 (UTC)

Good point. I have removed the block formatting.--Carl Hewitt 04:52, 3 November 2005 (UTC)

[edit] Actor model Prolog-like?

The article has the following sentence:

However, the Prolog-like concurrent programming languages (like the Actor model)

The actor model is not a programming language, and even if it were a programming language in what sense could it possibly be Prolog-like? --CSTAR 05:01, 3 November 2005 (UTC)

Good point. I have clarified the wording. Regards,--Carl Hewitt 05:09, 3 November 2005 (UTC)

[edit] Proposal to rename

Computation, mathematical logic and physics

The issues here have been raised by others (even by Feynman) and to pretend they are any way exclusively related to the actor model is silly.--CSTAR 04:00, 8 November 2005 (UTC)

The issues discussed in the article are related to the Actor model because it is based on message arrival orderings. The situation is more ambigous for the process calculi since I don't know of any published work. In order to generalize the title, some references and discussion to back it up will be needed in the article.--Carl Hewitt 06:17, 8 November 2005 (UTC)
For starters, how about
Leslie Lamport, Time, clocks and the ordering of events in a distributed system, Communications of the ACM, 21 (7) pp558-565
which I'm sure you know is one of the most important papers in the field.
--CSTAR 06:37, 8 November 2005 (UTC)
Lamport's paper is about Logical clocks and clock synchronization which is a different subject than what is discussed in this article.--Carl Hewitt 18:21, 8 November 2005 (UTC)
Because you say it's different? In fact, since it is unlikely anyone can even precisely say what this article Actor model, mathematical logic, and physics is about, how would it be possible to to exclude anything vaguely related to computation or physics.--CSTAR 18:49, 8 November 2005 (UTC)
How about renaming the article "Indeterminacy in computation"?--Carl Hewitt 19:06, 8 November 2005 (UTC)
That certainly seems like a more sensible choice. Before changing the name, I would hope other contributors would opine first.--CSTAR 19:12, 8 November 2005 (UTC)
No objection to that name from me. I still object to the POV physics references in the article, though. -- Arthur Rubin (talk) 22:41, 8 November 2005 (UTC)
Could you please explain what you mean by "POV physics reference".--CSTAR 23:23, 8 November 2005 (UTC)
"[[Arbiter (electronics)#Arbiters give rise to indeterminacy|physical indeterminacy]]" and "Physical indeterminacy in arbiters produces indeterminacy in Actors." It's not supported by physics. The other physics in the article, whether or not supported by fact, is in the context "The published claim....", so it isn't POV. Arthur Rubin (talk) 23:33, 8 November 2005 (UTC)
As far as I know, that's not POV. I would say it's an unsupported statement (and most likely a false one). Please, let's not confuse POV and falsehood.--CSTAR 23:36, 8 November 2005 (UTC)
PS. I misread what you wrote; I think those statements are basically OK. Arbiters do give rise to indeterminacy. other statements about the origin of indeterminacy in arbiters and arbiters being the origin of indeterminacy in distributed systems is what I object to.--23:42, CSTAR8 November 2005 (UTC)

[edit] New proposed title

The title proposed by Hewitt Indeterminacy in computation seems like a good one to me. Unfortunately, only Arthur Rubin has responded. If no one objects within the next few days, I will go ahead and change it.

The claim that indeterminacy may have physical causes appears to me to be an incontestable (or even trivial fact), so I don't understand Arthur Rubin's objections. I do object to the emphasis on the actor model, since as I mentioned the relation between physics and computation has been considered by several researchers (including von Neumann, Feynman, David Deutsch, Kitaev, Leslie Lamport etc.. besides Hewitt and some of his students). It may not be trivial to extract a useful article, however from their published work.--CSTAR 06:18, 10 November 2005 (UTC)

The claim that indeterminacy may have physical causes seems incontestable to me -- but this article is not effected by the cause of the indeterminacy. If the article were to be greatly expanded, then the question of whether the indeterminacy is physical, quantum, or just uncertainty as to the design might be relevant. Arthur Rubin (talk) 00:39, 11 November 2005 (UTC)
Agreed, what goes into the article may not be easy to decide.--CSTAR 01:30, 11 November 2005 (UTC)
I just realized that Indeterminacy in computation already exists and it's another one of the articles that was at the source of a prior dispute with Hewitt. Given that fact, I propose to merge them. Before doing so I will place a {{merge|Indeterminacy in computation}} banner on this article. --CSTAR 05:25, 12 November 2005 (UTC)
For what it is worth, I think this would improve things and give the reader who might not be that familiar with the topic a broader and more complete picture of the subject. DV8 2XL 11:19, 12 November 2005 (UTC)
Sorry for taking time in responding: I agree with the proposed merge. Even if the content of the article remains slanted towards CH's work, the article will still be better balanced if it gestures towards the context in the preamble and bibliography. It's worth noting that much of the relevant sources are widely available and fairly well known, so I'm hopeful about some of this editing work getting done. --- 16:49, 17 November 2005 (UTC)
I performed the merge into Indeterminacy in computation.--Carl Hewitt 23:18, 26 November 2005 (UTC)
Retrieved from "http://en.wikipedia.org../../../a/c/t/Talk%7EActor_model%2C_mathematical_logic%2C_and_physics_e867.html"