Wikipedia talk:Requests for arbitration/Carl Hewitt/Workshop

Discussion of Carl Hewitt's pattern of behavior

I am not comfortable editting the project page at this time, but here is how I see things.

Perhaps the best introduction to this issue is in a statement that Carl Hewitt chose to use as "Evidence": To myself (User:ems57fcva) from User:R.Koot:

Carl just started another vanity spree.

This is not a matter of creating vanity pages, but rather of creating content which gives the actor model and undeserved level of visibility and/or is very much focussed of Carl Hewitt's POV. When he goes too far, he finds himself up against the opposition of other editors. His responses are then to:

1. Try reverting the changes being imposed by other editors.
2. Move the offending content to a new location.
3. Move the offending content yet again.

As a result, editors find themselves having to fight a running, moving battle to get the content under control. The desire to have the actor model categorized under category:general relativity is just one example. It was quickly recognized by myself and other GR editors as software instead of theoretical physics, and the categorization was immediately bounced. Once Carl Hewitt saw that he could not do a direct categorization, he then tried an indirect categorization. When his initial term was ruled to be a neologism, he then tried to change the name (albeit incompetently, by creating a new category). As he sought an out, he would charge ahead with his new idea of where to put the content, without consulting other editors and without any regard to their insistance that the content itself is inappropriate.

This is the pattern that makes Carl Hewitt a very frustrating editor. I for one would not mind a 6-month ban on his editting in the article space. His writing is in general fairly measured, but he has no sense of how to work within a community of editors. He compromises only grudgingly, and often seems unable to draw the line between personal opinion and established viewpoints. --EMS | Talk 03:15, 8 December 2005 (UTC)

Response by Carl Hewitt

Please see Wikipedia:Requests for arbitration/Carl_Hewitt/Workshop#Calvinball, Lessons that can be learned from Edward Schaefer's pattern of behavior and How POV Neutral are Carl's_detractors? Regards, --Carl Hewitt 10:12, 6 January 2006 (UTC)

Lessons that can be learned from Edward Schaefer's pattern of behavior

Unfortunately, due in part to the limitations of our current communication technology, the Wikipedia can be a frustrating place to work. However, we press on anyway in building the world's largest free encyclopedia.

In the meantime, like me, Edward Schaefer is not perfect and there are some lessons that can be learned his pattern of behavior on the Wikipedia including the following.

Repetitively dragging up ancient history is a waste of time

Extracting from the other page

I have no doubt that in Carl Hewitt's mind he was attempting to preserve what he saw as valuable content, as opposed to maliciously evading the CfD. Yet the effect was the same. Simlarly, the contested content that was in quantum indeterminacy I have recently learned came from another article, and was placed afterwards in Actor model, mathematical logic, and physics. In response to an AfD attempt on the later resulting from the content being there, it got moved to incompleteness of quantum physics, which is actaully a good place for that content.

Admittedly, in hindsight, I could have done this more gracefully had I been a more experienced editor at the time. However, the content did end up in a more appropriate place. So we need to give it a rest.

The issue is an ongoing pattern of behavior, which is examplified by the above so-called ancient history, and which is quite visible in the very recent postings of yours below (which I intend to soon reference in the Evidence section, as it is part of said pattern). --EMS | Talk 16:32, 4 January 2006 (UTC)

Since this is the Wikipedia, nothing is ever finished, i.e. the interactions just keep on coming.

That said, please see Wikipedia:Requests for arbitration/Carl_Hewitt/Workshop#Calvinball, Attempting to win content negotiations by banning the other party will harm the Wikipedia and How POV Neutral are Carl's_detractors?

Regards, --Carl Hewitt 21:28, 6 January 2006 (UTC)

Attempting to win content negotiations by banning the other party will harm the Wikipedia

The complete quotation from which Edward extracted above is as follows:

• [1] Rudy Koot (aka Ruud Koot) to Edward Schaefer--Carl Hewitt 14:56, 6 December 2005 (UTC)
  • Carl just started another vanity spree. I want him block, quick. Any comments, suggestions or opinions? —R. Koot 03:51, 26 November 2005 (UTC)

Edward conveniently omitted the I want him block, quick. which goes with his later statement I for one would not mind a 6-month ban on his editting in the article space.

One lesson that I have drawn from my experience is that Wikipedia processes usually work if you give them time. However, banning editors who are trying to get the work done will prevent the Wikipedi processes from working.

Dispute resolution, however it may work out in this case, is an example of a Wikipedia process at work Fred Bauder 15:26, 4 January 2006 (UTC)

You have a good point. But of course my remark about was about ordinary Wikipedia processes involved in content negotation for articles, not extraordinary ones of banning Wikipedia contributors. My experience is that ordinary processes work if engaged in good faith and with persistence. Unfortunately, partly due to the limitations of our technology, these negotiations can be very difficult. Typically, I can settle a content negotiation issue with a colleague in 10 minutes face to face that might take weeks on the Wikipedia.

Sometimes another editor will not negotiate content. They might go into a mode of just complaining that they don't understand. I can empathise with this because it also happens to my students. Unfortunately, issues in Concurrent computing can get rather technical. Also Concurrent computing is currently a fast moving field that raises unresolved Wikipedia policy because articles in Category:Concurrent computing increasingly address isses that are on the edge of the state of the art. Without more expert contributors, I am afraid that the level of conflict and general frustration will increase. How can the Wikipedia address this policy issue?

In other cases an editor will become frustrated and drop out. I feel bad when this happens and hope that someday they will come back and resume negotiating no matter how frustrating our primitive technology. Others have become angry and called me names. When a Wikipedia editor found that their attempt to delete an article failed, they initiated this action to have me banned.

Of course this arbitration process will work. The question is what its effects will be. My colleagues have become increasingly concerned about developments on the Wikipedia. One of them told me, "If that is the way that it is, then I will have nothing to do with it!" However, the Wikipedia needs my colleagues in order to do a competent job in areas like Concurrent computing. So how can we attract them?

Regards, --Carl Hewitt 22:05, 4 January 2006 (UTC)

It takes a lot to drive people up a wall enough to initiate this kind of action. Enough work is involved with an action like this this that people do not care to go this route without compelling cause. In addition, there is also the fact that Carl is an expert in his field, and I agree that Wikipedia should not drive away people with such as that due to their potential to make significant positive contributions of Wikipedia. That is another reason why this action should not have been initiated without compelling cause.

So why was it initiated? Because there is compelling cause in the form of an ongoing an persistent patten of egregious behavior extending over the course of months, as documented by multiple editors and not just myself. When an editor continues over an extended period of time a disruptive and counterproductive pattern of behavior, it is reasonable to assume that said pattern will continue until and unless that editor if forced by strong action to chenge it. When the pattern is egregious enough, the fact that the person responsible for it is an expert matters none, since the disruption is the antithesis of the reason for wanting such a person to be present in Wikipedia. That is the issue here. --EMS | Talk 16:32, 4 January 2006 (UTC)

The amazing thing is how well the Wikipedia is working so far given what it is up against. However, past performance is not guaranteed to be duplicated as the Wikipedia continues its exponential growth. As Jimbo has said: We are making this up as we go along. Just trying to cope with our primitive technology ought to be sufficient to drive most people up the wall. Unfortuntately, it probably does because it compounds our problems with social issues including negotiating article content.

Unfortunately, a common human characteristic when it is difficult to work with others is to Blame the other person. This can be compounded by complaining loudly to the authorities causing a disruption which then also is blamed on the other person.

That said, neither I nor anyone else is perfect. (See Lessons that can be learned from Edward Schaefer's pattern of behavior, How POV Neutral are Carl's_detractors?, and Discussion of Carl_Hewitt's pattern of behavior.)

Regards, --Carl Hewitt 06:48, 6 January 2006 (UTC)

How POV Neutral are Carl's detractors?

The people out to get Carl are not, themselves POV neutral. They want to claim a strict hieracrhy of categories for physics topics and exclude anything that is not a physics result, i.e. not published in a peer-reviewed physics journal. I think that's too strict a standard. Carl is not claiming a physics result but a model of computation that crucially depends on physics results, i.e. a relation to physics that, I think, is important. His critics have not tried to understand it. It is deep and not easily understood. In the physicists' zeal to keep out crackpot science, I think, they've gone too far. Carl Hewitt is a reputable computer scientist [ Carl Hewitt ] that deserves to be argued with, not banned, and that the WP should want to retain.

Montalvo 03:16, 10 December 2005 (UTC)

Response to Montalvo by linas

There are multiple assertions here that I strongly object to.

• No one is "out to get" Carl. Carl has alienated over a dozen editors, as presented on the evidence page. Using language like "out to get" is inflammatory and is not helpful in this arbitration.
• "...are not, themselves POV neutral". There is no evidence of this claim, either on the evidence page, or elsewhere. Many of the complainants are people with advanced degrees, and specialists in a variety of fields, far more knowledgable than Carl. This is another highly disrespectful, inflammatory statement.
• "They want to claim a strict hieracrhy of categories for physics topics". This is not true. Physics is not a hierarchy, it is a collection of inter-related areas. None of the editors are claiming a hierarchy, and working physicists would not make such a claim.
• "exclude anything that is not a physics result, i.e. not published in a peer-reviewed physics journal." It is the (I believe unanimous) opinion of the editors in Wikipedia:WikiProject Mathematics and Wikipedia:WikiProject Physics that this is a core standard applied to math and physics articles in WP. For two reasons: 1) its quite easy to get drivel and crap published in peer reviewed journals, it happens all of the time. see e.g. the journal Foundations of Physics. 2) There is more math and physics published in peer-reviewed journals annually (every year) then there are articles in WP on all topics combined. WP deals with a microscopic portion of what is published. It is inapproriate for WP to start treating unpublished, and blatently incorrect, assertions.
• "His critics have not tried to understand it. It is deep and not easily understood." Carl's critics have PhD's in quantum mechanics and general relativity. Carl's critics understand these theories at levels far beyond what Carl understands.
• "Carl ... deserves to be argued with, not banned, and that the WP should want to retain." The evidence section of this arbitration directly contradicts this assertion. We all started with good intentions and a reasonable attitude. We tried conversing gently, and even arguing stridently, and none of these reasonable tactics worked. Carl is running totally out of control, and is unwilling or unable to engage in conventional norms of human behaviour. His presence here is disruptive to a degree far beyond the normal daily surge and ebb of WP disputes.

I am disappointed by this post, and note that such naive attitudes and statements by uninvolved parties are precisely the sort of thing that drives away able, intelligent and competent editors. We have some top-notch people here on WP: don't alienate them. linas 17:00, 23 December 2005 (UTC)

Linas - There is an undercurrent in your posting of "We are the good guys; You are the bad guys". I do not find that appropriate. For example are we "out to get" Carl Hewitt? I would say that in a sense we are, albeit with what we see as just and compelling cause. You wave around Ph.D.'s as if they mean something, but at this time the only Ph.D. involved in this action (to the best of my knowledge) is Carl. As much as we disagree with him, Montalvo is sincerely dealing what he sees as an injustice in this action, and he is also bitter about it. I wish that I could get him to see that Carl is operating in places and ways which are disruptive and counter-productive to Wikipedia, but I would not call him "naive" or anything else of that ilk for being concerned here. Do we really want to drive away people with good credentials like Carl's? Of course not. However, we want the types of behavior that he has shown in editting Wikipedia even less, and if he has helped to drive away other credentialed editors (such as CSTAR), then that is a reason for action. In short, I believe that we are on the right side here, but we aren't saints either. --EMS | Talk 03:10, 24 December 2005 (UTC)

I try to be as saintly as I can, all the time. I have a PhD in theoretical particle physics (that includes quantum, and a "minor" in gravitation and stuff like that). CSTAR has a PhD in physics (from Harvard); Charles Matthews and Jitse Niesen have PhD's in mathematics. Allen McInnes is in school, a PhD candidate in electrical/computer engineering. I assume Chris Hillman has a PhD, I don't know. Don't know about R.Koot. Sorry I'm slinging the PhD's; I'm simply trying to indicate that the other parties in the dispute should be assumed to be quite knowledgable. I didn't think that what Montalvo wrote is correct, in a number of ways, and I didn't want such assertions to stand unchallenged. I'm mostly trying to stay out of this, stay out of the way, sort of as a curious observer. I notice mostly that Carl has irritated a lot of people, drove a few to leave WP, and seems not to have altered his objectionable edit patterns one whit, and still doesn't "play nice" with others. linas 04:13, 24 December 2005 (UTC)

Sorry to have delayed in responding, but I have let my attention be elsewhere and had not noticed this earlier. I must admit to being most unimpressed by your Ph.D. as justification for your writings here. After all, Carl Hewitt himself can wave around not just a Ph.D. but probably also a long list of doctoral students whom he has helped to obtain theirs! If credentials alone where the issue here, Carl would be demanding that we back down, not the other way around. My concern is that you are reacting to Carl and his supporters on a very emotional level, such that is impairs your ability to make the case here. For example: I would say that we are "out to get" Carl, albeit with good reason. Instead, you are reacting to the negative insinuations of the term "out to get" above. Also, when you respond to the part on a "lack of a physical result" you start refering to "crap and drivel" instead of directly making the legitimate points that

1. unpublished work is usually baseless and never does achieve publication, and
2. that Wikipedia is not a crystal ball in that it cannot and should not attempt to predict what will be accepted in the future.

My request is that you speak to the facts and treat others here as your intellectual equals by default. After all, having a Ph.D. means that you successfully made your case for a thesis before a potentially unfriendly audience. Furthermore, you had to do so without putting down your detractors. All that I ask is that you use here the same thoughtfulness and tact that you would use before a the reviewers of your next thesis. --EMS | Talk 23:47, 5 January 2006 (UTC)

This is entirely inapropriate. I take your comments to be an outright attack on my character and personality. I resent the insinuations and I can only guess why you need to put me down. As to the comments about PhD's: You started this, not me. You stated that "at this time the only Ph.D. involved in this action (to the best of my knowledge) is Carl", and I corrected you. As to the phrase "out to get someone"; this is lynch-mob language, used to incite crowds. It is not appropriate for formal proceedings. As to my use of the phrase "drivel and crap", I apologize. Harrumph. linas 01:56, 6 January 2006 (UTC)

I apologize for hurting your feelings. Perhaps I should not have responded at all, but I honestly felt a need to express my own sincere concerns. It is a flaw of my own that I often fail to see the attack in comments like those of mine above, but instead was focussed on a desire to help prevent future conflicts. --EMS | Talk 04:29, 6 January 2006 (UTC)

Response to Montalvo by EMS

I won't claim a lack of POV in this matter, although I see my own view of things to be a resonably mainstream one. I understand the desire of some like yourself to defend Carl, but the actor model being dependent on physics does not make that physics dependent on the actor model, or even make the actor model relevant to the physics. Yet Carl has pushed that.

I see why Carl is seen as a valuable resource, but to me he is a persistent pain-in-the-neck. Perhaps I am especially annoyed with him since I do my own original research (see my user page) and have been scrupulous about keeping it out of the article space. Carl on the other hand has regularly interjected personal views all over the place, trying to stetch his "authority" into areas where his computer science expertise is not relevant. If anything, it makes me wonder if Wikipedia should not bar people from documenting their own work here. --EMS | Talk 06:02, 23 December 2005 (UTC)

Misstatements by Carl Hewitt

I am currently looking at a number of misstatements by Carl Hewitt:

1. It is being asserted that I am "trying to win the current negotiations by banning Carl Hewitt".
   1. What negotiations? I neither made any offers to Carl, nor have I received any from him. Without active, direct discussions between the parties on a given issue, there are no negotiations.
   2. I not trying to ban Carl here, but instead to restrict him.
2. It is being asserted that I want to ban editors from editting in "their area of experitse". That is not my intent in there actions. Instead I am saying that they should not be writing on their own work, and even that should be narrowly interpreted. The actor model is a creation of Carl's, and it should be described by others in order to have a neutral point of view. However, other models in computer science to which Carl has not made significant contributions, even if he has also discussed said models in his own works, can and should remain fair game in Wikipedia. I also explicitly make contributing to the discussions on the talk pages OK, since the comments from the creator of a theory or model are potentially invaluable to creating a better article. It is a matter of balance, and my suggestion may not reflect the best way of getting the most good while removing the most egregious misbehaviors. --EMS | Talk 19:55, 3 January 2006 (UTC)

---

The above discredited remarks which were added in the middle of my posting is just another example of Carl Hewitt's disruptive style of editing. See Wikipedia:Requests_for_arbitration/Carl_Hewitt/Workshop#Motion_by_user:ems57fcva for the responses and refutations. If find it amazing that Carl can respond to a posting about his misstatements by repeating his previous ones. Once again there is no acknowledgement of of this issues that I have raised, or even of the issues raised against these statements in the project page! This is very much evidence of the disruptive editting style that has resulted in this action, and all the more disturbing because it shows that Carl may have no clue as to the real reasons why this action is taking place. AS a result, he may prove totally unable to repond to this in a constructive manner. --EMS | Talk 05:49, 4 January 2006 (UTC)

Since Edward objected to my reposting of the issues from another page inline, I have removed them from above and have re-reposted them below. The reason that I have reposted them is that Edward stated above Without active, direct discussions between the parties on a given issue, there are no negotiations. So I need to get the issues back on the table since Edward decided to move the venue of the discussion from another page to here. It is true that Edward partially responded to some of these issues on the other page. However, I hesitate to move his remarks here. Of course Edward can move his responses here if he so desires.

Response to Edward Schaefer's motion which is quoted in italics below:

1) My motion is that Carl Hewitt be found to be a disruptive editor who regularly abuses the openness of Wikipedia to

• Promote the actor model in inappropriate categories and articles and/or to give it an undeserved importance.

The alledged unimportance of the Actor model is personal Original Research on the part of Edward Schaefer and is contrary to the published scientific literature (see Actor model).--Carl Hewitt 20:55, 22 December 2005 (UTC)

• Promote the Scientific Community Metaphor through inappropriate references in other articles. In general, the relationship of the scientific community metaphor to the subject in question is obscure and the mention undeserved. In addition, it is highly tied to Carl Hewitt's views of concurrent computing.

The alledged obscurity and undeservedness of the Scientific Community Metaphor is personal Original Research on the part of Edward Schaefer and is contrary to the published scientific literature (see Scientific Community Metaphor).--Carl Hewitt 20:55, 22 December 2005 (UTC)

• Promote in various articles on quantum mechanics such as quantum indeterminacy a view of the imcompleteness of quantum mechincs that is not generally accepted and which does not belong in mainstream quantum mechaics pages.

The article Incompleteness of quantum physics to which the Harvard physicist CSTAR and I have contributed reports on generally accepted published research of world-renowned physicists (Fuchs, Hawking, etc.) and consequently belongs in the mainstream pages of quantum physics.--Carl Hewitt 20:55, 22 December 2005 (UTC)

• These issues are exacerbated by Carl Hewitt's habit of regularly moving contested content of other articles or categories, forcing editors to chace [sic] down the content. Often this is done in the face of an AfD or CfD request, thereby evading the intent of the AfD or CfD. In the process, Carl Hewitt has caused much grief for other editors who have found themselves having to chace [sic] content and dealing with unproductive exchanges in talk pages on the contested content.

Edward Schaefer's comment above is objecting to the necessity for Wikipedia editors to sometimes reorganize the content of articles and create new articles in order to improve. In this course of this reorganization, we had to move content.

His allegations about AfDs and CfDs are myths.

Instead of engaging in the improvement of the Wikipedia, Edward prefers to try to ban those of us who are working.

Regards, --Carl Hewitt 11:17, 4 January 2006 (UTC)

---

The above reponses by Carl are nothing more than an attempt to divert the discussion from his own conduct, which is the reason for this action. I chose to move the discussion here because it was getting further and further afield from the desired findings of fact that it was supposed to be about. Now, having raised a question about his conduct on this page, Carl is trying to distract to conversation from that. I find this to be very disgusting and dishonest on his part. I also emphasize that this type of refusal to respond to the comments and concerns of others unless it should become absolutely necessary is one of the reasons whey Carl is dealing with this RfAr. --EMS | Talk 16:08, 4 January 2006 (UTC)

Remedies

In my opinion, a better remedy than being banned from autobiographical editing would be a mix of the following:

1. Carl should not edit physics articles (including subjects between computer science and physics) at all until his editing on computer science articles is judged to be generally constructive. Naturally he should be allowed to make suggestions on the talk page;

Yes, this struck me as a major problem. I don't much about either subject, but I do know they don't mix well. How could we phrase a remedy without regulating content? Fred Bauder 19:32, 4 February 2006 (UTC)

2. Carl should be encouraged to edit articles on subjects that do not arise from his research work (this suggestion was due to CSTAR, I think in one of the AfDs);

I'm sure he could improve many other articles. Fred Bauder 19:32, 4 February 2006 (UTC)

3. Carl's editing of articles on computer science related to his work should be covered by probation.

No, had enough of that stuff. Fred Bauder 19:32, 4 February 2006 (UTC)

There are three areas that I think that, if improved, would take a lot of pressure out of his content conflicts:

• Carl should try to aim his writing in articles at a less expert level, using only the most well known and least controversial concepts and formalisms within the field whenever possible;

This is true of many editors, Carl--please consider this. Fred Bauder 19:32, 4 February 2006 (UTC)

I completely agree that articles should be written to be as widely acessible as possible using the the least technically challenging and best established principles, concepts, and formalisms. Consider examples from the following articles:

• Actor model: The article is written to be accessible to anyone with a basic background in computer science. Its only use of formalism is a very simple use of the lambda calculus that should be readily understandable by almost anyone.
• Actor model theory: The article makes use of only the most basic properties of partial orders and is otherwise almost entirely self-contained.
• Denotational semantics: The article makes use of only the most basic properties of domain theory and fixed point semantics.

Of course the above Wikipedia articles need to actually cover the published literature on their subjects. Coverage can initiate reporting on principles, concepts, formalisms, and results that have been more recently published. Recent publications can be more accessible and less technically challenging than older ones. In addition they are more often available electronically in the Internet.

This is not to say that I have been perfect in observing the above principles and methods. Suggestions for improvement are greatly appreciated.

Regards, --Carl Hewitt 09:09, 5 February 2006 (UTC)

• While sometimes it is necessary to "cite deep" (ie. papers and books not available freely online, claims buried within articles that assume a lot of familiarity with nontrivial concepts that most CS graduates don't have), this should be heavily scaled back, and this should be done in discussions with editors as well as in articles;

Right Fred Bauder 19:32, 4 February 2006 (UTC)

I have been giving considerable thought on a development plan for references to address the following issues:

• Results from papers and books that are not freely available online At least for the Actor model, new papers will be published in the next year or so that that will make it unnecessary to refer to the older published work except for historical purposes. These new publications are needed in any case because of results that have been developed recently.
• Large numbers of reference for an article are often a symptom that the article is attempting to cover too much ground and that more specialized articles need to be written for subtopics. The more specialized articles will naturally have shorter reference lists. It is not clear to me that attempting to suppress a relevant reference does any good. Suppressing a reference from the Wikipedia does not make it less relevant. Wikipedia editors who know the literature are likely to mention it anyway if it is relevant to discussion about articles.
• Some concepts, principles, and formalisms are more advanced and technical is a considerable challenge for the Wikipedia. It can be addressed by making the head articles as minimally technically difficult as possible and providing a reasonable pedagogical path to more technically demanding leaf articles.

Regards, --Carl Hewitt 09:38, 5 February 2006 (UTC)

• Carl should make the habit of consistently citing the literature in Harvard style, so that claims are unambiguously linked to particular sources, likewise in discussions and articles.

We don't require any particular way of citing references Fred Bauder 19:32, 4 February 2006 (UTC)

I agree that it would be preferable to have a consistent way of citing articles and that the Harvard style is a good one. We should encourage its use. Is there any help available (templates, etc.)? --Carl Hewitt 09:58, 5 February 2006 (UTC)

I should note that I think that Carl's standard of engagement with other editors has improved since R Koot started putting together his original RfC, and I am optimistic that a well crafted ArbCom decision can mark an intermediate position that both allows editors who have had bad experiences with editing alongside him not to dread seeing him add new content, and encourages Carl to add good material on subjects that he is uniquely well placed to contribute.

I'd like, also, to see some kind of principle about how citations should be handled, since uncertainty about what to make to Carl's broad gestures at the extensive actor's literature has been a big factor in past conflicts. --- Charles Stewart 19:28, 12 January 2006 (UTC)

Citations need to associate the language in an article with specific language in the source. Fred Bauder 19:32, 4 February 2006 (UTC)

I agree with Fred that we need to do more of this. The tradeoff comes with the readability of articles. The exact relationship between the results that are reported in articles and the underlying scientific literature often is specialized material that more properly belongs in the article on the history of the subject. --Carl Hewitt 09:51, 5 February 2006 (UTC)

An interesting suggestion, and one that I am willing to try. I do counsel caution, however. As I note above, Carl still doesn't "get it" with our objections to his views in the area of physics. It also is unfortunate that someone as distinguished as Carl needs to hit over the head with a 4x4 to make him behave. If nothing else, your suggestion is missing a set of sanctions to impose as he gets out of line. (I once again emphasize that Carl is not acting with malicious intent, something which ironically makes the situation worse since this makes it difficult for him to see why he is earning the treatment that he is getting.) Yet I agree that Carl potentially can be an overall asset to Wikipedia, and that this medium is still figuring things out as it goes along.

Perhaps the thing that is a warning to us is in his questioning whether he can be effective or will be limited by the actions of the "mob". My experience is that when the "mob" is against you, there is a good reason for it, and it is your obligation to figure out what it is. Hapefully Carl will come to understand that obligation, and to accept it. It really, really is essential to being a constructive, productive member of the Wikipedia community. --EMS | Talk 20:29, 12 January 2006 (UTC)

The issues of mob psychology for the Wikipedia are tricky. Mob psychology is an important issue for my colleagues and one of the principle issues cited by many faculty for them not to contribute to the Wikipedia. Whether or not we can evolve effective procedures to address the issues raised by mob psychology will have important implications for the future of the Wikipedia. Some colleagues have drawn analogies to the case of J. Robert Oppenheimer. See

• Priscilla McMillan. The Ruin of J. Robert Oppenheimer and the Birth of the Modern Arms Race

The story of Oppenheimer's trial is a cautionary tale with lessons for us all.

Regards, --Carl Hewitt 01:32, 6 February 2006 (UTC)

I' very pleased you think this might be worth trying. One point: probation entails sanctions (probably progressively longer bans) were Carl to continue editing in the area (those computer science articles he has a stake in) in a disruptive manner, and I'm proposing that physics articles are out of bounds for some time. --- Charles Stewart 20:54, 12 January 2006 (UTC)

Sounds good to me. --EMS | Talk 21:38, 12 January 2006 (UTC)

This all sounds very good and I'd like to endorse these remidies. -- Koffieyahoo 08:14, 13 January 2006 (UTC)

Keeping this RfAr constructive

One thing that we should try to keep a focus on is that a good RfAr does not just propose remedies to prevent conflict, but also encourages editing activity that will help build a history of harmonious editing. Carl is enormously well placed to do this: the median theoretical computer science article is at a generally lousy level and Carl will have great experience in teaching computer science. If Carl were to choose a significant topic of interest to him but which is not of strategic importance for his research, then bringing this article up to a high standard would be a big accomplishment in this direction. --- Charles Stewart 21:48, 12 January 2006 (UTC)

I think you need to see what Carl is interested in doing. "You can lead a horse to water ...". If Carl gets constrained too much, he will simply bug out, and while some of us would not mind that in the least, it does not seem to be the intended result. Be advised that Carl is a theoretician, and sometimes great theoreticians make truly lousy teachers. So don't expect too much out of Carl, but above all do not start acting like you are his boss! He will, at dire need, work with us. What he will never do is work for us, nor do I think that he should. --EMS | Talk 03:51, 13 January 2006 (UTC)

Agreed (I'm a theoretician too!). Nothing along these lines should go into the RfAr unless Carl has indicated he thinks the idea is worth trying. --- Charles Stewart 03:59, 13 January 2006 (UTC)

Carl has his faults, but he is not a plain fool. If you wish to suggest areas that may be of interest to him, be my guest. IMO, the real issue is seeing to it that his edits are constuctive and appropriate, that he works with other editors instead of dominating a page with his own POV and frustrating those who seek a different approach. As I see it, those will be issues wherever he edits at this time. If he really is "mellowing" in the face of this action, then good. I too would like to see him continue here if he can be productive and constructive. However, I think that you are on sounder ground in the previous section of this discussion. Here it sounds like you are wondering where to put Carl, and I can tell you right now that you are not going to put him anywhere other than where he wants to be. I would not tolerate being constantly told what to edit, and I refuse to be a part of trying the impose any such thing on Carl. Appropriate restrictions on where he can go (such as "no physics editing") I am all in favor of, however. --EMS | Talk 05:43, 13 January 2006 (UTC)

In his Calvinball motion, he suggested concurrent computing as an area which WP needs improvement (and I agree: I don't find the treatment of the terms "concurrency", "parallelism" and "distributed computation" to be confidence inducing. Such an article I think would be too strategic to be appropriate, but I am sure that there are articles in this area that would be a good fit. --- Charles Stewart^(talk) 05:50, 13 January 2006 (UTC)

Just two small remarks:

• We should not force articles on Carl which we would like him to edit. Since this might trigger Carl into leaving, and that is not what we want.
• Given Carl's tract record in editing, e.g., the Parallel programming page, I would say that concurrent computing is too closely related to Carl's line of research. So, we should be cautious here. Off-topic: I'm still confused on how to reasonably make the subjects you mention above in a nice set of articles (see also the talk:parallel programming page. However, that shouldn't be discussed here, although I don't know where this should be discussed.

-- Koffieyahoo 08:26, 13 January 2006 (UTC)

Koffieyahoo - I'm glad that you also see my concern about forcing articles or editting areas on Carl. With regards to how to organize the pages, he may be of some help if he is approached respectfully on this. Personally I am not at all sure the it is possible to look respectful when doing so in the current context. IMO, this thread, in spite of it's title, is in danger of making this RfAr destructive. Wikipedia is not a sweatshop. I kindly suggest that we stop this awkward thread and its highly undesirable line of thinking here, and get back to the real issue of how to structure things so that Carl can potentially edit here while somehow ensuring that he will not get overly disruptive again. Otherwise, at least have the decency to ban him outright! Carl is a professor emeritus at MIT, not a dog on a leash! --EMS | Talk 14:37, 13 January 2006 (UTC)

I should say that I posted Carl with the intent to discuss this matter informally. --- Charles Stewart^(talk) 16:15, 13 January 2006 (UTC)

Motion by Charles Stewart concerning citations

Re this motion: I contacted SlimVirgin, and she said she might be able to think of something. --- Charles Stewart^(talk) 05:28, 13 January 2006 (UTC)

I agree to make the habit of consistently citing the literature in Harvard style if this is the general consensus. Moreover, I will agree to convert all of the references that I have contributed to the Harvard style (although it will take a while to do them all). My only concern is that if I contribute a reference to an article whose references are not currently in the Harvard style, which style should I use? --Carl Hewitt 06:45, 13 January 2006 (UTC)

Don't worry about past references: it's a shared task for cleanup on an article by article basis. Harvard style is recommended but not mandated by WP:CITE, and that reference explains how to cite tricky documents. --- Charles Stewart^(talk) 06:52, 13 January 2006 (UTC)

It is not clear to me what this motion is trying to achieve. I guess that "Harvard style" means something different to Charles, but for me, it refers to writing "F = ma (Newton, 1686)" instead of "F = ma [New86]" or "F = ma [12]". Was there a conflict related to the style of referencing? -- Jitse Niesen (talk) 14:18, 13 January 2006 (UTC)

No, not at all, the RfAr appropriate issue is finding a stable citation policy that reduces difficulties when verifying. I discussed citation styles with Carl a long time back, and he was agreeable to harvard style citations, which are the recommended style for WP:CITE --- Charles Stewart^(talk) 15:28, 13 January 2006 (UTC)

I believe that the issue is Carl's habit of generating lengthy "reference" lists, without making it clear which references are tied to which assertions. Requiring some form of direct citation would mitigate this issue. It would also make it easier for others to trim the reference lists down to a more reasonable length, since it would be obvious which references are directly related to the article in question. I think Charles is suggesting Harvard style as the specific type of direct citation simply because it is the recommended style. --Allan McInnes 18:48, 13 January 2006 (UTC)

Comment by User:CSTAR

It is with great unease that I venture into this dispute. After a period of inactivity (long by internet standards I suppose) I hope I don't sound too presumptuous.

Some of the proposed remedies don't seem to address the heart of the dispute: I think the most relevant comment in this dispute appears in User:Charles Matthews' statement Wikipedia:Requests for arbitration/Carl Hewitt

my reluctant conclusion is that this editor has a skewed view of what Wikipedia is here for; and that no amount of citations of policy will close down what is effectively a land-grab for undue prominence. Charles Matthews

The evidence for existence of this land grab is quite remarkable: Numerous articles in the sociology of science (Bruno Latour for instance; see here [2] for the relevant addition) as well as many articles in computer science. The fact of this land-grab should have been included in the evidence section, but infortunately is completely absent. To further employ this metaphor, on what justification is this "land grab" based? Published literature.

In other words, Hewitt has used the fact that some "publications" assert the connection between A and B and this fact alone justifies asserting the relation between A and B in Wikipedia. Again lots of evidence for this; for instance the whole sorry episode on arbiters, the actor model and their relation to quantum indeterminacy. This should be included in evidence.

Unfortunately, use of published literature in this way is a terrible idea. For instance, see Afshar experiment or Talk:Bell's theorem with which I am familiar. In both these instances, individuals with questionable or marginal (or even worse) proposals or theories in fact do have publications in refereed journals with their marginal (or worse) theories.

It was extremely frustrating for me to see Hewitt trot out (as if on a leash) the same long list of publications to justify any claim of his, whether it was capitalization of actor or the relevance of the actor model to relativity and non-determinism. This problem is not going to be fixed by some mechanical remedy such as adpoting Harvard style citations. Moreover, what about the converse: does the fact no reference exists in the published literature mean it is original research and should not be documented on WP? Well for one thing this poses an enormous disadvantage to documenting folk knowledge (for instance relating to African American or indigenous cultures.)

Moreover, what constitutes "published literature"? A seminar talk? A conference proceedings?

Big problem for Wikipedia, but that problem "s\~ao outros quinhientos cruzados"

Any remedy should try to accomplish the following: Hewitt must learn how to use the talk pages, and engage his collaborators more effectively. If Hewitt wants to play a constructive and influential role in WP, the way to do it is not by plastering references to the actor model all over Wikipedia, but by discussing his ideas with his fellow editors.

Let me suggest one standard I have tried to adhere to in my contribtions to Wikipedia The 1 revert rule. Except in very special cases (to wit: Intelligent design which is in constant struggle against creationists, in the quantum indeterminacy fiasco in support of User:ems57fcva's revert and in one particularly bizarre episode in the article on Ownership society are the only ones I can recall), I don't believe I have ever reverted more than once. Incredibly, I don't recall ever having reverted User:Caroline Thompson for instance! Although I did take a lot of grief.

A few other comments: Hewitt charges vandalism against both EMS and Ruud Koot. Well, he is not above a little vandalism himself.

05:30, 8 October 2005 Hewitt responds "Shoud we paraphrase Fuchs?-"[3]. Hewitt not only fails to answers my question regarding the statement on the actor model, he seems to have deleted it!

There are at least two other examples, which I don't have the patience to hunt down.

Unfortunately, I don't have any easy mechanical remedies to fix this problem; but I would like to remind people here this is a problem and that it is not going away. There is no official peer review system on WP and the informal consensus type review that was emerging was torpedoed by Hewitt's use of his reference "published literature."

Maybe my comments here should be moved to the Workshop section.

--CSTAR 17:20, 14 January 2006 (UTC)

Hello CSTAR!

Welcome back!!

With respect to your points

• I don't recall a single case in which Edward Schaefer has committed vandalism. In fact, he was a big help in dealing with another person's vandalism of the article Actor model. BTW, Ruud Koot seems to be calling himself Rudy again these days on his user page although he has recently signed as Ruud.
• With respect to your diff [4], I honestly don't know what happened. In some cases, strange things seem to happen to my own edits on the Wikipedia and I have to redo them. Gremlins? But I apologize anyway and would be pleased to answer any questions that you have about any of my contributions to the Wikipedia.
• Obviously it is insufficient to just to wave our hands and say the incantation "published literature" as justification in discuussions about articles. I must confess that have not been as specific as I should have been in the past about references and will do better.
• The Wikipedia may be the biggest intellecutal "land grab" in history. There are hundreds of thousands more articles now than when I started early last summer! Where did they all come from?
• With respect to synergies of the Scientific Community Metaphor with the work of Bruno Latour, Imre Lakatos, Tom Kuhn, Elihu M. Gerson, Susan Leigh Star, John Law, Geof Bowker, Michel Callon, Lucy Suchman, et. al., some of us are working to set up a session at the Society for Social Studies of Science (4S) Conference next year on this topic. It would be interesting to ask Bruno what he thinks about the statement [5] to which you took exception. He has previously expressed interest in the Scientific Community Metaphor.
• I like your idea of the 1 revert rule. The main purpose of revert should be to get editors to negotiate their differences on the discussion pages.

On another matter, I would like to see the article Incompleteness of quantum physics expanded to discuss relational quantum mechanics (in fact I have already put in a couple of references). Do you think that it would be better to put the expansion in the article or initially in the discussion page of the article? If it is placed on the discussion page and no one comments, how long should we wait before putting the material in the article?

Regards, --Carl Hewitt 04:14, 17 January 2006 (UTC)

Be careful, Carl. I will not accuse you of having knowingly engaged in vandalism, but some one your actions are akin to it (otherwise I would not be a complaintant in this action). Also do realize that the "vandalism" I helped you out with was actually an action done in good faith by an outraged editor. As a result, I have filed that incident under "stupidity" instead of "maliciousness/vandalism" (although I obviously do agree that the edit in question was totally inappropriate for reasons documented on the article's talk page). Kindly note that it was the last of that kind of incident from that editor (although he still does not like you), while I was engaged in regular battles with you up until this action was commenced. --EMS | Talk 05:33, 17 January 2006 (UTC)

I think that CSTAR's comment should be moved to the main case page as a statement by a complaining witness. --- Charles Stewart^(talk) 18:07, 16 January 2006 (UTC)

The use of indeterminacy

Carl Hewitt wrote in the project page:

The following is indeed a physical effect on which the Actor model depends: the indeterminacy of the fundamental partial ordering of the Actor model depends on physical indeterminacy. The reason is the use of Arbiters in the implementation of Actor systems (see Indeterminacy in computation). The consensus of the scientific community is that that Arbiters are physically indeterminate once they have become metastable. It is true that that there is a controversy about the source of the physical indeterminacy of Arbiters. Some experts maintain that it is simple physical indetermiacy of the usual kind with metastable systems. Others are not so sure. And some believe that the indeterminacy in Arbiters is due to quantum indeterminacy (see Incompleteness of quantum physics).

This does not validate treating the actor model or even arbiters as being a part of category:general relativity or category:quantum mechanics. The reason is because these things are users of the indeterminacy independent of its source. An analogy in computer science is Category:Internet protocols, which includes things like FTP, HTTP, IP, etc. However, it does not include things like Netscape and Internet Explorer, which depend on the internet protocols. The reason is because they are not concerned with the details of the protocols but instead just use them as black boxes. Similarly, aribters are not concerned with the details of quantum physics so much as they use it's effects to make a necessary decision. So arbiters are not a part of the theory of quantum mechanics (but they can be listed as as device which uses quantum mechanical effects directly). As for the actor model, it is a consumer of indeterminacy, and couldn't care less where the indeterminacy comes from. An indecisive user (or multiple users) and be just an indeterminate as an electronics circuit, after all. --EMS | Talk 18:34, 19 January 2006 (UTC)

For the purposes of computer science, the kind of indeterminacy can make a difference. For example, we distinguish between the nondeterminism of flipping a fair coin and the kind of indeterminacy of arbiters that is due to the necessary incompleteness of information that results from quantum physics. In fact we have constructed different mathematical models for nondeterminism and indeterminacy. Necessary incompleteness of information that results from quantum physics is an physical limitation for logic programming as opposed to a mere lack of information.--Carl Hewitt 11:00, 20 January 2006 (UTC)

Irrelevant as this does not speak to the point I am making. If I write an program that goes across the internet, it makes no difference from a categorization standpoint if it uses FTP or secure sockets or something else. The point is that the program consumes internet services instead of generating them. Similarly, I see no difference at the actor model level between the indeterminacy of a coin flip and that of an aribter choosing between simlultaneous messages. A football referee consumes the indeterminacy of a coin flip in just the same way as an arbiter consumes the indeterminacy of quantum physics, with the actor model then being a second-level consumer of the indeterminacy of the arbiter. Also, the aribter comes into play only for simultaneously arriving signals, while there are many other reasons why signals may be offset in time long before coming to the aribter, and be in either order thereby. You refuse to deal with that "classical uncertainty" even though it obviously is the major source of indeterminacy for the actor model. --EMS | Talk 16:54, 20 January 2006 (UTC)

There are several interesting cases to consider including the following

1. Classical physics of coins: Suppose that a coin is standing on its edge stationary on a large flat table. Now strike the coin perpendiculular to its face with a hammer that is much smaller than the coin at a point near the coin's edge where the tangent is perpendicular to the table. From knowledge the energy imparted by the hammer strike and knowledge of the twriling friction of the coin, we can bound how long before the coin settles. In this case we have bounded nondeterminism which corresponds to the classical theory of sequential computation, e.g. nondeterministic Turing machines. It is explicitly understood that (although extraordinarily unlikely) the result might always be "heads". Also by carefully observing the coin just after it has been struck, we may be able to better predict the outcome of "heads" or "tails".

   Once flipped, yes. Now try to predict how it is going to be flipped before it is flipped! --EMS | Talk 17:22, 24 January 2006 (UTC)

2. Quantum physics of Arbiters (controversial): Suppose that the same signal is symmetrically applied to both inputs of a symetric Arbiter and it goes metastable. The probablity that the Arbiter will not have stabilized grows expontially smaller in time. Consequently there is no bound that can be placed on how long the Arbiter will take to settle. I.e, the Arbiter always settles in a finite amount of time but the amount of time cannot be bounded in advance. In this case, we have unbounded nondeterminism which corresponds to some theories of concurrency e.g., the Actor model. Also, carefully observing the Arbiter in order to try to bound the time that it takes to settle only serves to move the indeterminacy elsewhere and does not bound the time. Consequently carefully observing the arbiter just after the two signals have been input can never enable us to better predict which of the two possible stable states will result. Also in this case there are fundamenatl physical limitations on what it is possible to compute using Logic programming (see Indeterminacy in computation).

It should be obvious from the above discussion that the circumstances for models of computation are quite different with the classical physics of coins as opposed to the quantum physics of arbiters.

Regards, Carl Hewitt 15:33 23 January 2006 (PST)

You have not demonstrated that in the least. You don't know which state an aribter will settle into until is settles. You don't know which side a coin will land on until it has landed. You don't know which clerk will send their message first. The cause is different in each case, but the indeterminacy is the same. Even if it is not the same, the actor model is still a consumer of indeterminacy instead of a creator, an issue that you have chosen to ignore. --EMS | Talk 20:20, 23 January 2006 (UTC)

Leslie Lamport gives a very good account of how unbounded nondeterminism can arise in decision problems in his paper on Buridan's principle. The canonical example is "Buridan's Ass" - the ass that starves because it is placed equidistant between two piles of hay, and can't decide which one to eat from before it starves. Buridan's principle implies that unbounded nondeterminism is simply the result of trying to make a discrete decision based on a continuous input. There is no need to resort to quantum indeterminacy to explain the the unbounded nature of the nondeterminism. In fact, Lamport points out that, since physics at the quantum level is not continuous, some people might argue that it would be possible to avoid unbounded nondeterminism through the use of a quantum mechanical arbiter. He then presents a reasonably convincing argument (but not a formal proof) that even at the quantum level unbounded nondeterminism is probably unavoidable. Bottom line: unbounded nondeterminism is not the result of quantum indeterminacy, but a simple consequence of Buridan's principle. --Allan McInnes 22:58, 23 January 2006 (UTC)

Welcome to the controversy! As far as I have been able to determine, the consensus of expert opinion is that the quantum physics of arbiters may not be needed, but if it is then it could strengthen the results and provide additional insight.

For example as stated in Buridan's principle, "Similarly, we cannot conclude from Buridan's Principle that the decision will take an unbounded time." Similarly from Buridan's Principle alone, we cannot conclude any of the following results that are important for indeterminacy in the Actor model:

• The probability that a metastable Arbiter has not settled declines exponentially with time.
• Increasingly precise measurments of an Arbiter just as it has entered metastability will not enable us to better predict which stable state will result.

Bottom line: Buridan's Principle by itself is not sufficient to derive the properties of indeterminacy for Actors.

Regards, Carl Hewitt 17:04 23 January 2006 (PST)

You have taken the quote from Lamport's paper out of context. Lamport states that

Buridan's Principle rests upon mathematical concepts of continuity and boundedness that are not physically observable. No real experiment, having finite precision, can demonstrate the presence or absence of continuity, which is defined in terms of limits. No experiment can demonstrate that an arbiter requires an unbounded length of time to reach a decision. An experiment in which the arbiter failed to decide within a week does not prove that it would not always decide within a year.

He then draws an analogy between Kepler's laws and Buridan's principle, to illustrate the fact Buridan's principle represents a theoretical ideal which practical finite measurements will tend to be consistent with.

Furthermore:

• The fact that an arbiter circuit has an exponentially decreasing probability of remaining metastable is a function of circuit design (and can be - as far as I am aware - adequately derived through standard small signal analysis).
• The metastability properties of an arbiter circuit are a theoretical ideal, just as Buridan's principle is, since no finite experiment can confirm them.
• The definition of metastability effectively requires that it is impossible to determine which stable state will eventually result - otherwise the system wouldn't be metastable.
• Do you have any references for this "expert consensus" about the applicability of quantum mechanics to understanding arbiters? (I note there are no such references at Arbiter (electronics), even though you appear to be the principal author of that article - perhaps you could add some specific references that back up this "consensus" there).
• What aspects of "the properties of unbounded nondeterminism for Actors" are not covered by Buridan's principle? I don't see any mention of exponentially decaying probabilities or increasingly precise measurements in Will Clinger's dissertation. Are you suggesting that Clinger's denotational model doesn't fully capture the semantics of Actors?
• The fact remains that the existence of unbounded nondeterminism does not rely on quantum indeterminacy.

--Allan McInnes 02:35, 24 January 2006 (UTC)

The subject here is properties of indeterminacy in the Actor model.

My information about the opinions of experts comes from my asking them about the question. This often what happens at the edge of the state of the art ;-) So unfortunately, nothing is published yet. But hold on, that should change fairly soon.

Do you know of any published proofs of the following: the probability of a metastable Arbiter not having stabilized decreases exponentially in time?

Also do you agree with the following principle:

• Increasingly precise measurments of an Arbiter just as it has entered metastability do not enable better predictions which stable state will result.

If you agree with the above principle, how do you suppose that it could be derived?

Regards, Carl Hewitt 19:39 23 January 2006 (PST)

Yes, the subject is properties of nondeterminism (or indeterminacy if you prefer) in the Actor model. Which is, as far as I can tell, a mathematical theory that does not rely on the specific metastable probability characteristics of an arbiter circuit (although it does rely on the assumption that some kind of arbiter can order message arrivals). Or are you claiming that if someone was able to produce (a) an arbiter with a non-exponential rate of stabilization, or (b) somehow created an metastability-free arbiter, the Actor model would no longer be valid?

Perhaps it would be best to wait until this state-of-the-art material you refer to is published before trying to disseminate these views on Wikipedia. I say this not because of any concerns about original research (although that is a concern), but because without any references to back up your claims the inevitable result is the running arguments that appear to have surrounded this issue since it was first introduced here.

Regarding published proofs of exponential stabilization probabilities, you might try "Synchronisation and arbitration circuits in digital systems".

I have no particular inclination one way or another on your proposed principle. For a start, it depends on what you mean by metastability (i.e. the actual metastable saddle point, or just the close neighbourhood of that point). Secondly, it depends on what assumptions you make about measurement — see "Buridan's Law of Measurement" in Lamport's paper. Thirdly, you haven't cited any empirical evidence to support the principle, and I am not aware of any supporting evidence either.

--Allan McInnes 07:18, 24 January 2006 (UTC)

Since, nondeterminism has an already well established meaning in computer science, it is important for us to focus here (see section title) on properties of indeterminacy for the Actor model in order avoid confusion.

The issue that concurrency is a fast moving field is unlikely to change anytime in the forseeble future. So waiting for the dust to settle may not be the best option for the Wikipedia. The situation is fairly simple for the Wikipedia articles: only report on published work. But what do we do in the talk pages where we are trying to sort thngs out and information on the current streams of unpublished research (talks at conferences, preprints, etc.) can be highly relevant?

Properties of indeterminacy enter into the Actor model in various ways including the following principles:

1. The following property of Actor indeterminacy can be important for calculating how long an Actor computation is likely to run: the probability of a metastable Arbiter not having stabilized decreases exponentially in time.
2. The following property of Actor indeterminacy can be important for showing that Actor computations are more powerful than those of logic programming: Increasingly precise measurment of an Arbiter just as it has entered metastability do not enable better prediction which stable state will result. Consequently logic programming cannot in general deduce an Actor computation based on increasingly precise observation beforehand.

The second principle above can be expanded as follows: It is realatively easy to guarantee that a symetric Arbiter enters metastability: just provide the same signal symetrically to both inputs. Currently our best theory of measurement is that of quantum physics and so, of course, we should use it.

There is a large published literature that backs up the second principle above. So the question remains on the table: From what physical theory could the principle possibly be derived?

Regards, Carl Hewitt 01:49 24 January 2006 (PST)

Red herring again. To quote from the above:

From what physical theory could the principle [of arbiter non-determinacy] possibly be derived?

That only makes my point that the actor model consumes indeterminacy, and so is not relevant to any source of the that indeterminacy such as quantum mechanics. (BTW - general relativity, in which you tried to place the actor model and so got myself aware of you, is not a source of indeterminacy in that it is a classical theory anyway!) --EMS | Talk 17:22, 24 January 2006 (UTC)

Sigh. This will (I hope) be my last contribution to this particular thread. It seems silly to carry on an argument of this nature, particularly given the venue in which it is being carried out. With that said, I would like to make the following comments:

1. What, precisely, are the properties of "Actor indeterminacy" that differ from unbounded nondeterminism? Is it the notion of an exponential decay in probability of staying metastable? Do you really want to tie the Actor model to one specific model of arbiter (which is only an ideal). I thought the Actor model was more general than that. Is it the idea that more precise measurements won't help with predicting outcomes? I fail to see how that differs from traditional unbounded nondeterminism, which essentially says that there is no way to predict the outcome.
2. The essence of science is reasoned argument. It's hard to have a reasoned argument on a talk page, becuase there isn't sufficient room. That's why you see continued requests for specific publications (journal, conference, preprint, or otherwise) that explain the theories you are propounding: the requestors are looking for some kind of external reference that does have sufficient room to lay out a reasoned argument, and allows them to evaluate that argument on its merits. Without references you are making unfounded assertions and vague appeals to authority. Not surprisingly, arguments ensue.
3. A calculation of how long an Actor computation will take obviously depends on the time it takes for arbiter decisions to resolve. However, that should be a parameter of the calculation, shouldn't it? Different arbiters have different rates of exponential decay. Not to mention the fact that the exponential model itself is just that: a model (based on a linear analysis).
4. Theory and experiment would seem to suggest that better measurements of metastability inducing initial conditions can provide a better prediction of which stable state will result. This is hardly surprising: there is only one truly metastable point in the phase space of an arbiter, so more precise measurements (and good characterization of the phase space) should permit the detection and prediction of "near metastable trajectories". A truly metastable state has been proven "almost surely" impossible in Newtonian arbiters. Quantum arbiters may raise different questions, but such arbiters are pretty much just theory as far as I know.
5. The ability to predict an eventual outcome is different from the "detection" of metastability. The latter boils down to yet another decision problem on continuous inputs (are the inputs in the "metastability inducing region" or not), which is itself subject to metastability.
6. Could you please provide specific citations for the "large published literature" that supports your "second principle"?

--Allan McInnes 18:00, 24 January 2006 (UTC)

Bottom line: quantum physics applies to Arbiters. What is needed at this point is a detailed analysis of the behavior of Arbiters in terms of quantum physics.

The paper "Synchronization Circuit Performance" by Kinniment, Bystrov, and Kakovlev does not claim to enable better prediction of Arbiter outcomes from initial measuments while entering metastability.

In the published literature it is agreed that the probability that a well-designed metastable Arbiter has not stabilized declines expontentially with time. Such Arbiters always settle in a finite but unbounded time. Claims that such Arbiters are "impossible" are contrary to the accepted scientific literature (even if they seem to appear in "Proving Newtonian Arbiters Correct, Almost Surely" by Ian Mitchell).

The Actor model must be sufficiently flexible to make use of the above information about Arbiters. The limitation on the generality of the Actor model is that it must follow the laws of physics. In this respect the Actor model differs from some other models of computation.

Remaining stuck in the old "nondeterministic" computational paradigm with it concepts, principles, and methods is not going to advance the cause just as remaining stuck in the concepts, principles, and methods of classical physics will not advance the cause. A paradigm shift is required.

Regards, --Carl Hewitt 11:10, 30 January 2006 (UTC)

Quantum physics applies to every physical system. However, not every phenomenon requires a quantum-level explanation. You have yet to cite a source for your claim that quantum physics is required to describe the behavior of an arbiter. I have cited several sources that demonstrate that arbiters, unless explicitly designed to utilize quantum mechanical effects, can be described in a classical framework.

I said that the results in Kinniment, Bystrov, and Kakovlev suggested that better measurements would yield better predictions. Specifically, from their conclusions: For metastability times of 8τ or more, well within the normal operating range of a synchronizer, the output becomes non-deterministic, and neither the final value nor the individual output time depend primarily on the inputs. Which, to me at least, suggests that for times less than 8τ the outputs are deterministic. You may, of course, disagree.

Mitchell does not claim that arbiters will not settle in finite time. In fact, the point of his work is to prove that they always will settle, because it is "almost sure" that the arbiter will never actually find itself in a truly (as opposed to apparent) metastable state. There is nothing contrary to the literature in his claims. In fact the proofs of exponential probability of settling that I have seen (and cited above) appear to rely on the existence of trajectories that present the appearance of metastability, but do not actually intersect the true metastable saddle point, as their basis.

You assert that a paradigm shift is required. That may be true (although I have yet to see you present any evidence to suggest that classical models are inadequate for arbiters). However, unless there is published literature on this paradigm shift (which, again, you have not provided citations for) such a paradigm shift is necesarily original research, and does not belong on Wikipedia.

Enough. Further debate on this issue is completely pointless without literature demonstrating that arbiters actually do need to be analyzed in terms of quantum physics. --Allan McInnes 17:43, 30 January 2006 (UTC)

It is true that our discussions have reached the edge of the state of the art. Furthermore, discussions of Wikipedia articles in Category:Concurrent computing stand to increasingly come up against issues that are at the edge of the state of the art. One of the aspects of the edge of the state of the art is that there is relevant information available (preprints, talks at conferences, etc.) that has not been published. Information that has not been published clearly cannot be reported in Wikipedia articles. However such information can be highly relevant in discussions about Wikipedia articles. It would be good to have some policy guidance on how to conduct such discussions.

Demonstrating that Arbiters can be described by classical physics alone requires demonstrating that all of their physical properties can be predicted classically. So far this has not been done in the literature. Mixing in classical assumptions at the start of the discussion does not help (such as being in a "true" as opposed to an "apparent" classical state and the "appearance" of being in a classical state).

When the same input is provided to both inputs of a symmetric Arbiter then it is certain that it will become "truly" metastable and not just "appear" to be metastable or "apparently" become metastable. Furthermore the outcome is absolutely indeterminate regardless of how long it takes the Arbiter to settle. There is nothing "almost sure" about the Arbiter settling: it is certain that it will settle. Better measurements of such an Arbiter entering metastability do not enable better predictions.

1. Please don't shout.
2. Discussing "relevant information" (preprints, talks at conferences, etc.) is only possible if you provide citations or links to that relevant information so that others can familiarize themselves with it. You have consistently refused to do this, both here and at Talk:Arbiter (electronics), which makes discussion impossible.
3. What properties of arbiters are not adequately described by classical models?
   1. "Overview of Modelling and Analysis Techniques for Arbiters and Related Circuits" gives a comprehensive overview of the dynamical systems approach to modeling arbiters, and makes no mention of any properties not adequately represented by classical models
   2. "Synchronisation and arbitration circuits in digital systems" shows how the exponential probability of stabilization can be derived classically
   3. "Measurements on a High Speed Arbiter" demonstrate empirically that real arbiters match the classically-based physical models
   4. "Metastable States in Asynchronous Digital Systems: Avoidable or Unavoidable?" shows that metastability is a general phenomenon which arises wherever a continuous set of states has to be mapped onto a discrete one, that metastability cannot be avoided in principle with classical devices (i.e. metastability is a property of classical systems), that quantum devices can in principle avoid metastability (i.e. metastability is not necessarily a property of quantum mechanical systems), and that other quantum effects like macroscopic quantum tunneling prevent a practical application of such quantum devices (i.e. even quantum mechanical systems will suffer from metastability - which is not the same as saying that metastability is caused by quantum effects).
4. The only hits that Google returns on quantum indeterminacy and arbiter metastability appear to be Wikipedia articles that you have written, or mirrors of those articles elsewhere. The same appears to be true of quantum decoherence and arbiter metastability. Obviously not definitive, but disconcerting nonetheless.

--Allan McInnes 23:43, 30 January 2006 (UTC)

The issues we are discussing are right at the edge of the state of the art. Furthermore they are controversial. Several experts share your opinion that quantum physics may not be needed. But they seem to lack convincing argumeents that classical physics is entirely sufficient. Consider the following properties:

1. Exponential Settling Property: The probability that a metastable Arbiter has not settled declines exponentially with time. Clearly a derivation of the Exponential Settling Property from quantum physics could be very convincing. But a derivation from classical physics is not so convincing because how can we know that quantum phenomena are not at work in the Arbiter?
2. More Precise Initial Measurement Doesen't Help Preditability Property: More precisely measuring an Arbiter as it is becoming metastable does not provide for better prediction which outcome will result. It seems that deriving the More Precise Initial Measurement Doesen't Help Preditability Property requires quantum physics.

Since the issues are at the edge of the state of the art, they are now being discussed in personal conversations, talks at conferences, preprints, referee reports, panel discussions, etc. What do you propose that we do about this nonpublished information in disucssions on the Wikipedia?

In due course there will be more published papers on these issues that can be reported in Wikipedia articles. Of course, even after the next article has been published the issues may still be controversial. For example, suppose that I were to publish an article that that claims that quantum phenomena are involved in the operation of Arbiters? We could report the article on the Wikipedia as one view, but it wouldn't settle the controversies;-)

Regards, --Carl Hewitt 01:45, 31 January 2006 (UTC)

1. The classical derivation of the exponential settling property has been well-verified by experiment. It does not matter if there are quantum phenomena at work, unless those phenomena have an impact on the behavior of the arbiter. There are "quantum phenomena" at work in everything, but many things can be perfectly adequately modelled by classical techniques. Any claim that metastability is caused by some quantum phenomenon requires showing that the existing classical explanations are inadequate. Until there are experiments that demonstrate non-classical behavior in an arbiter, and work that demonstrates that some hypothesis based on quantum phenomena explains that non-classical behavior, there is no need to resort to a quantum explanation. That's how science works.
2. Your More Precise Initial Measurement Doesen't Help Preditability Property is (a) a theoretical ideality, rather than a proven physical property of real arbiters (unless you can cite otherwise), and (b) does not necessarily require a quantum-mechanical explanation, since such behavior could be produced by sensitive dependence on initial conditions in a classical dynamical system. Again, to justify discarding the existing classical models it is necessary to show that the classical models are not able to account for the observed physical behavior of real arbiters.
3. As to how to handle the discussion of completely unpublished material (i.e. material that is not available in any printed or electronic form), I suggest that it is pointless to try to have a discussion about it - no one else can join the discussion, since they have not been exposed to the material in question. Where there is material that has not been officially published in a journal, but can be made available as a preprint or a conference paper, discussions are possible. However, as it stands, you have not provided any preprints, any conference papers, any pointers to conference programs that show papers covering these hypotheses, or even named any researchers doing work in this area. So discussion is, as I have stated several times, pretty much impossible. All we are left with your assertions. --Allan McInnes 03:44, 31 January 2006 (UTC)

Sigh. I probably should not add to this argument. However, both of Carl's remarks are easily rebutted. In real-life physics, when a formula can be derived by using only classical physics, then this is taken to be sufficient evidence that quantum mechanics does not play a role. This is not considered to be a sign of weakness of the theory; on the contrary, its a sign of strength and correctness. As to the second point about predictablity and accuracy of measurement, no quantum mechanics is needed. It is well known that chaotic systems have a positive Lyapunov exponent, which means that "more precise measurements" won't give you better prediction. A standing example is predicting the weather. Prosaically, the phenomenon is known as the Butterfly effect, and does not rquire qunatum mechanics for its explanation. linas 02:38, 31 January 2006 (UTC)

Linas,

Welcome to the controversy!

In physics, it is useful to know exactly what approximations are being made. So it would be ideal if we could show that a classical analysis of the behavior of Arbiters is an appropriate approximation of a quantum physics analysis. Nothing wrong with that! However without a quantum physics analysis, it is natural to wonder what approximations are being made and in what circumstances they are appropriate.

It is true that the outcome of an Arbiter operation can be extremely sensitive to its initial conditions. Also since the operation can take unbounded long, more precise initial measurement does not necessarily enable more accurate prediction of outcome even in the classical case. Nevertheless a quantum physics analysis of Arbiter behavior could resolve questions about the nature of Arbiter indeterminacy. For example it could address the question as to whether there is necessarily more indeterminacy than would be predicted by classical physics alone.

Regards, --Carl Hewitt 04:13, 31 January 2006 (UTC)

I knew I should not have opened my mouth. Carl, you make several incorrect statements here. First, as to "it is natural to wonder what approximations are being made": the approximations that are made in solid state physics are well-understood and are commonly taught in undergraduate and graduate physics courses. Thus, at best, your wondering is vague and unfocused, at worst, misguided.

As to "...address the question as to whether there is necessarily more indeterminacy than would be predicted by classical physics alone." This statement is also incorrect. A laboratory experiment, comparing a theoretical prediction to an actual measurement, would resolve (or prolong) any supposed controversy.

My complaint with these types of statements is that they are vague, misleading, inaccurate, pseudo-scientific POV pushing, based on a misunderstanding of what physics is and how it works. Its the same language and tone that the inventors of UFO propulsion devices use. Its inappropriate and wrong for WP articles. linas 16:26, 31 January 2006 (UTC)

Hi Linas,

It's nice to see that we are going to keep this discussion on track not degenerate into personal attacks ;-)

Yes, indeed, topics of the kind being discussed here are studied in solid state physics where various approximations are used in quantum physics calculations. The problem is that that none of the experts whom I have consulted know of any publications on the quantum solid state physics of Arbiters. Do you know of any?

Regards, Carl

If the experts don't know, then why did you write an article purporting to contain the answer? And why to you continue to pursue this line of argumentation? linas 03:11, 1 February 2006 (UTC)

Edward Schaefer keeps bringing up the same things because he seems to hold some resentments about long ago content negotations. See his comments in this section and elsewhere in this action.

The area we are discussing is controversial and experts have not yet reached agreement. Recently I have learned a couple of things that I didn't know. The published work is only indirectly applicable and more work remains to be done. Hopefully there will soon be more publications.

Regards, --Carl Hewitt 04:54, 1 February 2006 (UTC)

Best argument concerning this issue I've heard in months. -- Koffieyahoo 07:59, 20 January 2006 (UTC)

Fine, but this has nothing to do with either quantum mehanics or with general relativity. Its precisely these kinds of vague, misleading, incorrect, inappropriate, uncited, unverifiable claims about quantum mechanics and general relativity that paved the way to this RfA, and I am disheartened to see a reappearence of this style of argumentation. linas 19:46, 23 January 2006 (UTC)

I'm not all that disheartened. It is nice the Carl is showing his true colors here. This makes it obvious that if he is allowed to stay in Wikipedia, then he must be kept on a short leash. --EMS | Talk 20:20, 23 January 2006 (UTC)

Proposed remedy revisited

I'm optimistic about my proposed remedy above providing a workable solution to this RfAr, because of EMS's support, and because in an offline discussion with Carl, while he had reservations about the proposal, he was not hostile to it. I'd like to tackle some issues now:

1. One of Carl's reservations was how to define what are the scope of computer science and physics. It seems that the wikiprojects might help to define this, although the computer science wikiproject is only now, through Allan McInnes' efforts being revived. An issue here is how to determine what to do with those few articles in the overlap. I suggest we can let discussions of the project talk pages determine how to handle these issues if they arise;

2. Possible EMS has second thoughts after so much time has past?

3. I haven't talked to Ruud about the proposal. As the third party to this RfAr, I'd value his input.

--- Charles Stewart^(talk) 21:38, 3 February 2006 (UTC)

Which proposed remedy are you advancing? Fred Bauder 12:19, 4 February 2006 (UTC)

In #Remedies. --- Charles Stewart^(talk) 19:20, 4 February 2006 (UTC)

---

I remain leary of Carl after all that he has put us through, and noting that

[o]ne of Carl's reservations was how to define what are the scope of computer science and physics.

is a red flag that more trouble is to come if something is not done. I find it hard to believe that someone of his intellect could not see that line if he wanted to. Just to repeat: The actor model is a consumer of indeteminacy, and is not concerned with its source. Instead it the Carl Hewitt that likes to be concerned with that. The arbiter that Carl loves to cite exists at the boundary, being an electonics circuit that generates digital indeterminacy while consuming quantum indeterminacy in the process. How an arbiter operates is electronics (with the details of electronics being the physics behind it), but how an arbiter is used is computer science. Since "electronics" intervenes between computer science and physics, I see that issue of an overlap as being a red herring.

With the passage of time, I become more and more convinced that being a expert in a field and being a good editor about the field are two very different things. A good editor can collect information from a variety of sources and produce a comprehensive and comprehendable article on a subject from them. An expert in a field, espectially one who is actively expanding the knowledge of that field is instead often more concerned about esoteric details, and often ends up writing narrowly focussed and jargon-laden articles. Sometimes, the experts are totally POV, or so full of themselves that they cannot write an properly encyclopedic article on any subject that they care about. My advice is to ignore Carl's status as an expert, and focus on Carl as an editor. On the one hand, he is adding value to Wikipedia through sharing his knowledge, while on the other he is being (ar at least has been) disruptive and counter-productive. The issue before us is one of whether the benefit is worth the bother. --EMS | Talk 19:31, 6 February 2006 (UTC)

For me, I think the benefit outweighs the bother, but then I am more interested in the computer science articles than the physics articles, I am optimistic about Carl editing in a more constructive manner, and I regard this as a test case for highly accomplished expert editors that has something substantial at stake. Part of the reason I proposed the phased remedy is that it is clear that for most editors of physics articles, the bother much outweighs the potential benefit. --- Charles Stewart^(talk) 22:32, 7 February 2006 (UTC)