Identity
Filter
Visitors
(1) (0) (0) (0) (0) (1) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (61) (0) (0) (0)Subscribers
(0) (0)
Highest posts
- 08/11 21:28 +1 Re: Lemming-DP
- 06/11 04:33 +1 Re: Lemming-DP
- 05/11 01:09 +1 Re: Lemming-DP
- 04/11 21:56 +1 Re: Lemming-DP
- 04/11 01:20 +1 Re: DP-based-collectives instead of FA/DP
- 04/11 00:22 +1 Lemming-DP
- 02/11 21:16 +1 Re: DP-based-collectives instead of FA/DP
- 02/11 07:14 +1 Re: DP-based-collectives instead of FA/DP
- 01/11 21:59 +1 Re: DP-based-collectives instead of FA/DP
- 01/11 07:00 +1 Re: DP-based-collectives instead of FA/DP
Latest posts
- 08/11 21:28 +1 Re: Lemming-DP
- 06/11 04:33 +1 Re: Lemming-DP
- 05/11 01:09 +1 Re: Lemming-DP
- 04/11 21:56 +1 Re: Lemming-DP
- 04/11 01:20 +1 Re: DP-based-collectives instead of FA/DP
- 04/11 00:22 +1 Lemming-DP
- 02/11 21:16 +1 Re: DP-based-collectives instead of FA/DP
- 02/11 07:14 +1 Re: DP-based-collectives instead of FA/DP
- 01/11 21:59 +1 Re: DP-based-collectives instead of FA/DP
- 01/11 07:00 +1 Re: DP-based-collectives instead of FA/DP
Links
Blogs
Democracy
Search
[top-politics] Re: DP-based-collectives instead of FA/DP
-
At 08:22 PM 10/27/2006, you wrote:
> >L: Range Voting is a voting technique designed to pick an optimized > > winner, in terms of the sum of social satisfaction with the result. > > The immediate result, that is, the election itself. It is no > > guarantee that the voters will still be happy when they look back. > >-M: Agreed. >By contrast, with SD2-S, the goal is to find those that are best at >knowing future satisfaction – to eliminate the populistic >*quick-fix*.
The FA/DP equivalent is the proxy system. I consider it very likely that high-level proxies will be, on average, these people. FA/DP is not likely to come up with stupid, quick-fix advice.
Now, suppose we have a set of such people. Are they all automatically in agreement? If so, fine, no problem?
But when I said that humans were good at collective estimation, I meant that the average guess of a large number of people with unbiased information (such as a jar they can actually look at full of pennies, with the job being to guess how many pennies there are in it) will be quite good. At least this is the claim of the “wisdom of crowds” people, and I think it is true. This is not lemming-wisdom, unless the lemmings are using range voting to determine the common movement.
Range is an averaging method. The rating of an outcome, when Range is used, is the average rating of the voters (the mean, though some Range methods discard outliers, which moves toward the median).
If you can find that subset of voters who are more skilled — there is discussion about how to do that — then you would use Range to analyze their ratings of possible choices. If you can’t define the subset without introducing bias, then you use the entire electorate.
Proxy methods weight votes by proxy counts, so if there is Range Voting by proxy, one is precisely weighting the votes of proxies according to how widely they are trusted. This is using the human capacity for determining trust, for choosing whom to trust.
Again, this faculty is quite good on average, when the clients have direct experience of the ones being trusted. By no means is it infallible, people can be fooled even when they know the “candidates” face to face, even when they know them over time and apparently deeply. But all that is necessary is that when a voter chooses a proxy, the proxy is on average as trustworthy as the voter or more trustworthy. If this condition is met, then delegable proxy will amplify trustworthiness as one moves up the proxy tree. On average. Which if Range is used as a voting method, an average improvement in the trustworthiness of the voters will result in an average improvement in th trustworthiness of the poll result.
DP/Range could be used in a power structure. But if it is used in an FA/DP organization, which is a pure communications structure, there is the additional protective layer of the decentralization of power. The point is that FA/DP advice should generally be trustworthy, and that members have a direct means of knowing that. They are not relying on some computer system to tell them that they ought to trust the results, they are trusting an individual whom they have chosen, who advises them. And they also have, if they wish, the overall poll results, they can tell if what their proxy is telling them is a general consensus or is idiosyncratic.
> >L: However, humans are actually pretty good at making estimates, if they > > have unbiased information (not necessarily very much information, but > > what they have has not been slanted), and if they can express their > > estimates in a refined way. Not just Yes or No. > >-M: Where-the-flip is this ‘unbiased’ information going to come from? >Sounds like wishful thinking. :-(
Good question. Combined with a hyper-reactive comment. I did not claim that they would have such information, I was commenting on what happens if they do. So, indeed, unbiased information becomes very important. How does one get this information?
Well, one of the most common ways is to pay for it. When we hire a professional to advise us, we expect the advice to be unbiased. Sure, it is not always so, but, if it is not, we have chosen our advisor badly.
So if we, the public, want unbiased news, we have to pay for it, I’d suggest. If we let someone else pay for it — which is mostly the system we have — then we are quite likely to get what the payer wants us to get. We will get what information benefits or agrees with the goals of the one paying for it.
This is why it has become quite important, with scientific studies, that the funding sources be disclosed. That a funding source might benefit from acceptance of the study results does not mean that the study was biased, but it does mean that we should be aware of the possibility. Drug companies benefit from studies which show that their proprietary drugs are helpful, so they pay for those studies. Will they pay for a study which tests whether or not their drug is superior to, say, a dietary change? Maybe, but not usually. It is too risky for them. These kind of studies come from public institutions, or, less often, from food companies. And since you can’t patent foods, generally, the economic motivation for such research isn’t there.
So if we want better information, to repeat, we need to pay for it. How are we going to do that? There are two basic ways. We can identify existing professionals whom we can hire, or companies with whom we can contract, or we can create new companies. We may also be able to use our own proxy network as a news-gathering service, but I’m only noting that as a third possibility, and one which might not work in the long run, for reasons I won’t go into, except to note that this involves volunteer labor, and, for a system to be stable, long-term, there are special conditions which must be satisfied, depending on volunteers is tricky. We can get away with in the FA/DP network, to a degree, but gathering news is work, and it requires skill and dedication. And it deserves compensation.
So aside from what could develop out of the chaotic structures of an FA/DP organization — which are theoretically unbiased — we have quite the same problem as we have with politics: how can we coordinate and cooperate?
I assume that a political FA/DP organization will spawn media initiatives. They might be as simple as advising members to subscribe to certain existing media, to support existing unbiased media. There are media professionals dedicated to unbiased reporting and analysis. Once again, FA/DP organizations don’t endorse outside organizations, that is part of the FA traditions. But, once again, FA caucuses can. They do so internally, not as representatives of the FA, which they are not. However, if a consensus has appeared in the FA, the result would be as if the FA had made a single recommendation.
FAs are rigorously unbiased. The only bias in an FA is the collective bias of the members; but I also expect that functional DP networks will understand the bias problem, will understand that bias leads to bad decisions. (If it did not, it would not be bias, it would be intelligence and knowledge, or perhaps excellent intuition.)
> >L: I don’t think anyone here (i.e., following this list) has any good > > idea of how Range Voting would function in a political context. Mark > > does not know what he says. He’s making predictions and stating > > them as if they were facts. > >-M: I know how easily the lemmings are duped. >Range-voting is just a more sophisticated way to measure this duping.
It is quite true that Range used with uninformed voters will not produce informed results. Range does not magically turn a sow’s ear into a silk purse. A collection of pig’s ears is still a collection of pig’s ears.
The average opinion of lemmings is, quite simply, the average opinion of the lead lemmings.
However, if those leaders are informed….
DP, as Mark should know, is a hybrid system. It allows voters the freedom to vote directly, but most voters will quite wisely, I expect, delegate their votes to those more directly involved with the issues, those with more knowledge, more time, or, yes, more intelligence, more ability to understand.
DP produces the desirable results of healthy oligarchical structures without imposing an oligarchy.
And Range Voting is only a polling method. It does not take the place of deliberative process. My own opinion is that deliberative process is best as a binary tree, and that is precisely the Robert’s Rules method. Every question is Yes/No. A motion evolves through a series of Yes/No votes until it is passed — or rejected. There is no half-acceptance of a motion. (Though there is tabling, i.e., let’s consider this later, when we have more time, more information, or have gathered more support or opposition.)
>-M: Again, what is reqired for good decision making is ‘unbiased’ >information. >Do you see this happening anytime soon, anywhere?
Yes. If.
>Until then, the glaze-eyed lemmings will be obediantly following their >puppet-masters and spin-doctors.
Yes. If.
>This is why I keep insisting on having a multi-order c-algorithm to >filter away the lemmings.
Oh, Great Mark, if only we will trust you, all will be well. We have to trust you, because you are not presenting us information in a manner we can understand.
I have said again and again that filtering is essential to any mass intelligence. It appears that Mark is presenting us with an engineering solution. We probably are not going to understand it until we see it functioning.
The same problem exists to some extent with FA/DP, but the latter, if people will spend a very little time with it, is extremely simple. The elements are all tested and known to work, except perhaps delegable proxy — but the proxy concept is thoroughly established and it is amazing that it is never considered when it comes to politics. It is ubiquitous with economic and other rights. Delegability is a small but important twist that makes proxy structures scalable. I think I can see some of the implications, but with any engineering solution, there will be hosts of unanticipated consequences.
Still, the obstacle to understanding FA/DP are mostly unexamined but ubiquitous assumptions about what is possible. Direct democracy, for example, is impossible on a large scale. We think. Because we have never considered delegable proxy. (This group is an exception, delegable proxy is quite well accepted here as a general concept.)
The FA concept is quite controversial here, but not among people with experience with FAs. For the latter, what is new is the possibility of applying FA principles outside the narrow definitions of existing FAs. But quite a few have seen this coming, I’m not the first person to think of it.
Where I am first, I think, is in seeing what could happen if FA traditions were married to delegable proxy. FAs really live at the small meeting level, they function only in a minimal way on a large scale, AA does it through supermajority-elected delegates; when a supermajority cannot be obtained, they use a coin toss among the top two candidates. The idea is to broaden representation beyond mere majority selection. It works, apparently. But I think we can do better.
FA/DP does not require elections at all….
>-M: Thats because there is nothing to not work. >Communicate and poll away while the world passes by.
Here is what Mark’s comment boils down to: Don’t talk, don’t negotiate, just follow what our computer tells you should be done, based on inputs you provide.
What if the poll is “Shall we advise members to fund the campaign of Joe Shmoe?”
An FA/DP can conduct such a poll. The results do not send, directly, any money to Joe Shmoe, because the FA doesn’t collect money and distribute it like that. However, proxies might have expenditure authority, individually granted by their clients. There might be a banking structure where members can effectively vote with a bank account. Other clients might reserve all expenditure for themselves. This is an utterly free choice in an FA. You can give your proxy the right to spend funds from an account you have dedicated for that purpose, the only choice you make, then, is how much to put in that account. And your proxy could delegate this. So you could end up with individuals with large resources at their immediate disposal, all voluntarily contributed.
So the proxy structure passes the recommendation back down. That is, those proxies who support it pass it down. The others might be passing down a recomendation to send money to the opponent in that campaign.
Economic balance provides strong incentive for the FA to discover consensus, for, if it does not, if action is recommended prior to finding consensus, much of what is spent will essentially be wasted, because it will be opposed. FAs, in a sense, vote with the member’s pocketbooks, which the members control.
>-M: No ‘it’ does NOT advise, because there is no collective voice to >advise. >*IT* is just a VOICELESS communication network, therefore it is >wussified and weak.
The human brain is thus weak, because there are only individual neurons firing. What a terrible design! One set of neurons may be attempting to move the organism in this direction, another in the opposite direction. The organism will be paralyzed.
No, it will never fly, my analysis proves it. Heavier than air, impossible.
Unbiased communication is inherently “weak.” Coercion is bias.
>By contrast, SD2-S would provide THREE voices: >1. The generalist-trustee voice >2. The specialist voice >(this is the voice that is organizationally decisive) >3. The popular voice >(this is the voice that must be appeased to get a decision past >deliberation.)
All these voices are present in FA/DP, and identifiable as such. It does not have to be complicated.
All this yak from Mark, and I have no idea, really, what SD2-S is. The name does not appear to be descriptive. FA/DP, Free Association/Delegable Proxy, is descriptive, though Free Association has been given a specific meaning, it is most easily described by reference to the Alcoholics Anonymous Traditions as generalized. Delegable Proxy will be readily understood by many simply from the name, I’ve experienced the reaction simply by saying it, there are a few people who will say “Aha! What a great idea!” One of them immediately said, “What about loops?” — showing that he was paying attention.
(Loops are a non-problem presenting themselves as a difficulty requiring some kind of organizational fix. The most that an organization has to do about proxy loops is to inform the members when they are not represented in a vote. It is up to them to fix it. Systems that force a fix are in danger of introducing bias or possibly corruption.)
> > >-M: Its a communication structure with polling. > > >L: Yup. But the kind of communication structure that we have never seen > > formalized. Cellular, robust, difficult to centrally control, > > efficiently filtered with minimal loss of information, fully > > democratic and non-coercive. > >-M: Information about proxy chains is guaranteed to be lost unless one >uses a multi-order c-algorithm like PageRank. Yet so many systems want >to ignore the proxy’s proxy.
Actually, if the loop infrastructure is informationally antibiased, the intended unwindings won’t disentangle random proxy caucuses.
:-)
> > >L: I have asked you about the input field, but you seem to DODGE this. > > >What would the input field be like? > > >L: You’d better explain what you mean. “Input field” means to me > > something that someone fills out in a form and submits. “The input > > field” for what? The proxy structure is simple,… > >-M: Yes, proxy inputs. This needs an input field.
Well, duh! What, exactly, has not been said about this that needs to be said. Basic proxy assignment database:
Client Member Handle, Proxy Member Handle, Acceptance by Named Proxy (Yes/No), (Comment)
>-M: WHERE? Where is LOMAX’S input field? :-( >You gave GENERAL GUIDELINES, but I don’t see any input field.
Oh! Let me guess: Mark is expecting that we have set up a system where a member can interact with a database to enter the proxy assignment information. We have not done this for any organization. Demoex did it, I think.
There are proxy assignment lists that have been set up. So far, they are only editable databases. There is no interface, users just add their record to the database. If they don’t follow the suggested form, other users may edit their record to take it into the proper form, but editing to alter an indicated proxy assignment would be considered fraud, I presume it might be sanctioned, though FAs don’t punish members.
When it is needed, I assume that we will set up more sophisticated systems. An easy one is a mail-in form that appends an emailed record to a wiki page. This can be enabled in TikiWiki by clicking on some boxes…. just haven’t done it. Nobody has requested that BeyondPolitics set up a proxy list. When you only have two or three active members, it is not yet needed.
(This may not be true. I really should set it up. There is a broader group of people who have expressed interest, but who are not active in any way at the present time. Let me see, what we need to involve these people is … delegable proxy…. But the fact is that I can only do so much, and I’m refraining, making a virtue of necessity, from doing more than a certain amount of work on the actual structure. BeyondPolitics.org is not intended to be the mass political organization that we foresee. It is just about FA/DP as a method. It’s theoretically focused, intending to provide support to the actual organizations that will function politically and in other fields.)
> >L: That is correct. They are their members, nothing more. However, > > membership in an FA is open. Membership is self-declared. If not, if > > membership is controlled, they are not FAs. […] > > The two together have never existed. We don’t really know what would > > happen. But I think I can see some of it. I seriously think it is > > worth trying. There were times in my life where I could simply have > > implemented it. I’m not in that position now, and it would take quite > > a bit of work to get back there. I’ve decided not to do that, not to > > try to reach a position of power, but to focus on the theory and on > > education and the coordination of attempts. > > And thus my participation here. Some of the work here is reasonably > > aligned with FA/DP concepts. Some of the tools developed may be useful. > >-M: FA/DP has polling. >This means that it will have both an input field and an output field.
Marks language is obscure. What is “it.” FA/DP is an organizational concept. It may use polls, and, of course, polls have “input fields.” There is no output field, necessarily, the poll consists of a collection of records, each record being one vote. The poll does not generate a result that says “The motion passed.” That report, if true, is made by a human being who uses analysis of the two databases involved: the poll results, which consists of votes, perhaps in the form “Member Handle, [Vote],” where [Vote} for a standard Robert’s Rules motion is Yes, No, Abstain, and then there is the proxy database.
Thus votes may be reported in a number of ways. They can be reported as the raw vote numbers. 12 members voted Yes and 3 voted No. Or the votes can be expanded by use of the proxy list. That same vote, expanded might be, 23 members, voting directly or by proxy, voted Yes, and 3,487 members, voting directly or by proxy, voted No.
Both votes are of interest, aren’t they? To really make sense of these votes requires more information than is present in the two databases. We can tell from the numbers I gave that members voting directly favored the measure, but there were many more members whose proxies — at least one proxy — voted in opposition. What happens from this vote?
That is up to the caucuses involved. The FA generally is only measuring consensus. The caucuses will decide whether or not to act.
The caucuses can and will use additional information. For example, the numbers above don’t show substantial support for Yes. You have at most 12 members who favor the motion, enough that they took the trouble of voting for it. One wonders, indeed, how this motion even got to an actual vote….
But the No votes, what do they mean? They could mean that a few high-level proxies agree on No. Or they could mean that one person has marshalled a sock-puppet army. The caucuses will be able to tell by examining the identity of the voters and their general behavior in the organization. If they are unknowns, the caucuses will be completely free to discount those votes, all those votes might be only one person, whose position should be noted but not necessarily followed….
>What will people SEE that will make your system so amazing? >I even proposed an output field, but you seemed to dodge this.
What output field? Mark is obscure in what he wants, and then blames the people he is communicating with for not providing the information he wants. Dysfunctional.
>-M: Yes, you are supposed to understand EVERYTHING that I say, or >request clarification.
Sorry. I’m not available. I’m already married.
>As for Markov chains, in our context this refers to proxy chains when >proxies themselves have proxies. Difficult? Obscure?
Yes. Why the term Markov chains? What does it add to “proxy chains”? Delegable proxy tells us that there will be proxy chains.
>Lomax, have you tried understanding PageRank/SD2 at all? If not, maybe >SD2-S is the miracle system you have been looking for all along, but >you have been ignoring it.
I’m not looking for a miracle system. I’ve been relying on Marks criticism of FA/DP as information that SD2-S does not satisfy the FA/DP conditions. It may well be DP.
Mark expects the world to beat a path to his door. He bought that mousetrap myth. Build a better mousetrap and starve, unless you learn how to market it or engage someone to market it for you.
Mark, we are lemmings! Get it?
The mice can be crawling all over the people, but, not having a mousetrap, the people have become so accustomed to them that they don’t even notice them any more. Show them your mousetrap and they won’t see it. What’s that? Why should we put this ugly thing in our houses? Isn’t that dangerous?
No, if you care about your system, you are going to have to learn to move lemmings. As they are, not as you want them to be. Same is true for me. I’m working on it. I think I know how, but it takes time. It is really the only problem.
>As for SD2-S, all proxies are required to endorse other proxies.
Yeah, we considered that. I rejected it. If you don’t assign a proxy, this is your prerogative. If you choose a proxy who does not assign a proxy, this is likewise your prerogative. All that is necessary is that people know — if they want to know — that they are unrepresented. That is easy to determine. For any poll, any member who did not vote and who did not have a proxy, regardless of chain length, voting, can be informed. The member can shut down the information process, though a better solution for the member might be to name a proxy.
It has been proposed that default proxies be assigned. Maybe. It creates, I think, more problems than it solves. Not every member need be involved in every decision, not even indirectly. What is important is that the member may be involved if the member chooses to be involved. And that the member has a means of delegating involvement. Not a requirement. That’s what I’ve come to so far. Actual organizations, of course, will make their own decisions about this, and BeyondPolitics.org will be very interested in how it works out. We hope that people will tell us.
>This creates the potential for extremely long MARKOV CHAINS, as well as >the potential for LOOPS, and creates forks. Problematic?
No.
>Not for the voter, nor for the analyst if PageRank is being used, which >is why SD2-S uses PageRank.
And which I don’t understand at all, and which I’m not highly motivated to attempt to understand. Because I don’t see the problem for the analyst in the first place. If a member is not represented, so?
The FA context makes a lot of precaution that is necessary in power structures unnecessary. However, even in power structures, I don’t see a problem. I don’t want to automate member participation, I want participation to be a free choice. Participate or leave it to others. If you want, you can choose who the others are. If you don’t, well, that is your prerogative, it is no skin off our teeth. Participate whenever it suits you, that is what suits us.
> >L: Yes. But I’m not going to argue the point. If Mark is right, why, of > > course, all we have to do is to go be selected by an SD2-S based Wiki. > > But what is going to create that wiki? > >-M: Maybe Markus. He has asked me to provide specifications. >I am still thinking.
So you have nothing. You are just thinking, which is even less than communicating. Sauce for the goose, Mark.
Creating the wiki is still not creating the organization. There are a number of nearly empty attempts around the internet. I think of Mikael Nordfors — one of the independent inventors of delegable proxy, I think — and Nornia.org. I just tried to look at nornia.org but it did not respond. Last time I looked, Nornia.org was a few people having made a few comments. The intention was to set it up as a “delegated voting” organization — he does not use the term “proxy,” he calls proxies “advisors.”
You can open the doors, but unless people come in, you don’t have a party. You’ve got some open doors.
> >L: Where is one? Show me! > >-M: I think that I have provided more specifics for SD2-S than you have >for specific DP systems.
In other words, what has been provided is theory, which is what I’ve said I’ve been providing. There is no “input field” for SD2-S, there is a plan to create one, some day. Yet Mark behaves as if what is being compared is a clearly established system, and what others are proposing are mere theories.
Dysfunctional, again.
I do not recall the specifics Mark is talking about. If this is a coherent theory, more than what has been provided by me, here and on beyondpolitics.org — and what is here is generally more developed — then where is it expressed? Where is the system design?
Mark used “Markov chains” without defining the term. I looked it up, which is more than most people will do. The definition did not inform me at all of what he meant by it.
> > > > In AA, the central functions are handled by AAWS, Inc. But AAWS was > > > > designed to hold very little property, the minimum necessary for its > > > > function plus a “prudent reserve,” in Bill Wilson’s terminology. Why? > > > >-M: I want SD2-S to be the best form of collective action possible, > > >if so, I want as much property organized under it as possible. > > >*Why* do you seem to set your standards low? > > >L: Because what Mark wants has been tried so many times. Collecting > > large resources for central control attracts parasites. They will > > figure out a way to corrupt the system. > >-M: Will – this pisses me off.
Good. I like pissing off arrogant pretenders. It’s a fault of mine. I apologize.
>You are assuming that SD2-S is corruptable simply because it is >centralized. >Yet you don’t understand it, do you?
No. I’m basing this in an inference. It could be wrong. Show me. I do have some ideas about what SD2-S is, but I have not seen sufficient definition to claim to understand it.
> >L: The safeguard against this: the decentralization of power… > >-M: Then it no longer has the benefits of centalization. >Lets cure or headaches by cutting off our heads.
No, FA/DP does create a centralized structure, but it does it as a fractal. Smoke that, Mark. If a human brain — which is fractally organized — is “centralized,” then FA/DP could be said to be centralized. But there is no “center.” There is a gestalt that appears, we call it “identity.” Technically, any Buddhist would know that it does not exist, it is an illusion. But decisions get made anyway, as if there were a central identity.
>-M: Then you have the potential problems of centralization there. >How about this?: >DO CENTRALIZATION BETTER!
Centralization can be done better. But there is something even better, which is centralization in coordinating structures with ultimate control being radically decentralized. This decentralization protects the centralized intelligence.
Anhone who thinks FA/DP is simply not centralized does not understand it. Aspects are centralized and aspects are decentralized. Power is decentralized. But communication is centralized, not in the sense of being centrally controlled, but in the sense of consensus being continually discovered. Centralization is powerful because it mimics consensus. Consensus is more powerful than centralization, because it is unity that is not based on coercion. So it takes less energy to maintain the unity, that is why it is more efficient and thus more powerful. But consensus process, traditionally, is difficult, inefficient. Delegable Proxy is designed to make it efficient. That’s what I expect. We won’t know for sure until it is actually applied in large-scale organizations. But first it must be applied in small ones. Unless we get lucky. Which I’m not counting on.
(Someone in a position of power in a large organization may decide to implement a DP structure. There was a time when I could have done it. But people in such positions usually don’t see the need. After all, the existing system succeeded brilliantly because it put them in their position of power, right? If it ain’t broke, don’t fix it.)
> > >-M: Lomax, you can be so annoying. :-( > > >L: Mark, you can be so obtuse. I’d rather be annoying. > >-M: Markov chain c-algorithms are easy to understand.
Wow. I must really be an idiot! Presumably they are so easy to understand that cogent explanation isn’t necessary. Except for idiots and lemmings.
>-M: With SD2-S, there would probably be a treasury committee composed >of high ranking adminstrators, and whose policy has been determined by >treasury related specialists.
Lots of ways to do it.
In FA/DP there is effectively no central treasury. However, caucuses can set up organizations which function in any way they choose. One of the interesting concepts would be what’s been called the House of Money. It is called the House because it is like the House of Representatives, except what is represented is member funding. Members put funds into an account, and they can “vote” with these funds. And they can delegate their “vote.” Which would mean that high-level proxies could quickly marshal and dedicate very significant funds. However, this is not FA. By definition. It is, however, quite possible as an independent structure, and it could be DP. Structures like this would be the muscles of the body of which FA/DP would be the intelligence.
> >L: Can the majority decide to spend the money in the central treasury? > >-M: No, but their wishes would be known to the top ranked specialists >for the given issue. >I predict that a reasonable compromise would be made.
So the specialists can spend the money, i.e., a majority of them?
Look, Mark, I think you are making this up as you are going along. I can understand that. To some extent, I’m doing that as well. But I’m not claiming that “the house of money,” for example, is part of the FA/DP concept. It is one way that it could work out. I think this, just what you have said:
I predict that a reasonable compromise would be made.
I am not fixing the specific ways in which FA/DP organizations would function. They will work that out. I’m laying out general principles, and sometimes, because people ask, showing how these principles could possibly be applied. The general principles are the Free Association concept, which creates a rigorously neutral, noncoercive organization, avoiding even the subtle coercion of “dues or fees” and central treasuries, and the Delegable Proxy concept, which is really just a scalable extension of standard free proxies, including free assignment. Proxies are not elected, they are chosen. Nobody can force you to choose a proxy, a coerced proxy is not a proxy but a recipient of stolen goods. You don’t have to compete with anyone in choosing a proxy. It is utterly different from elected representation.
> >L: Against the wishes of some of those who contributed it? > >-M: No system will appease all the people all the time.
In other words, yes. “Appease” betrays an attitude, possible. He could have written “please.”
What Mark is describing is standard organizational practice. FAs do not do this in any significant way, and the concept is a proven one. It works. Organizational unity is enhanced, which is the purpose of avoiding this kind of centralization.
> >L: If so, you have created a disincentive for people to join who might > > fear that they will be funding something they oppose…. > > Congratulations. Welcome to same old same old. > >-M: Yes, the same old of centralization via democracy. >There are no other options.
I’ve been describing one with all these words, for all this time. It seems Mark hasn’t heard any of it. Maybe someone has, now or in the future.
Otherwise, why bother?
+1
-
04/11/06 00:22
Lemming-DP
Lomax wrote: […] >L: Mark expects the world to beat a path to his door. He bought that > mousetrap myth. Build a better mousetrap and starve, unless you learn > how to market it or engage someone to market it for you. > Mark, we are lemmings! Get it?
-M: We are systems designers. I expect there to be few lemmings among us.
And I expect most of us to understand the basics of SNA and Markov-based centrality algorithms.>L: The mice can be crawling all over the people, but, not having a > mousetrap, the people have become so accustomed to them that they > don’t even notice them any more. Show them your mousetrap and they > won’t see it. What’s that? Why should we put this ugly thing in our > houses? Isn’t that dangerous? No, if you care about your system, you are going to have to learn to move lemmings.
-M: No, I only need to move a few non-lemmings.
>L: As they are, not as you want them to be. Same is true for me. I’m working on it. I think I know how, but it takes time. It is really the only problem.
-M: Maybe you should have a DP-based macro programmed for Excel and/or OpenOffice.
People are more sensate than theoretical, and want to see something work.> >L:As for SD2-S, all proxies are required to endorse other proxies.
>L: Yeah, we considered that. I rejected it. If you don’t assign a proxy, > this is your prerogative. If you choose a proxy who does not assign a > proxy, this is likewise your prerogative.
-M: That makes their data less PageRankable. Markov algorithms work by following these chains.
With SD2-S, the person still has the prerogative to not vote for proxies, but this is just turning the vote over to the computer, who will vote for underdogs.>L: All that is necessary is that people know — if they want to know — that they are > unrepresented. That is easy to determine. For any poll, any member > who did not vote and who did not have a proxy, regardless of chain > length, voting, can be informed. The member can shut down the > information process, though a better solution for the member might be > to name a proxy.
-M: My way, everyone is represented, always.
>L: It has been proposed that default proxies be assigned. Maybe. It > creates, I think, more problems than it solves. Not every member need > be involved in every decision, not even indirectly. What is important > is that the member may be involved if the member chooses to be > involved. And that the member has a means of delegating involvement. > Not a requirement. That’s what I’ve come to so far. Actual > organizations, of course, will make their own decisions about this, > and BeyondPolitics.org will be very interested in how it works out. > We hope that people will tell us.
-M: SD2-S has:
- default general proxies (to be discouraged – the computer selects underdogs.)
- general proxies(trustees), who are general organizers, and serve as default specialist proxies.
- specialist proxies(specialists), are the selected authorities for specific issues.
> >This creates the potential for extremely long MARKOV CHAINS, as well as > >the potential for LOOPS, and creates forks. Problematic? Not for the voter, nor for the analyst if PageRank is being used, which is why SD2-S uses PageRank.
>L: And which I don’t understand at all, and which I’m not highly > motivated to attempt to understand. Because I don’t see the problem > for the analyst in the first place. If a member is not represented, so?
-M: Then they aren’t providing useful proxy information. The voter is the data input for the c-algorithm.
>L: The FA context makes a lot of precaution that is necessary in power > structures unnecessary. However, even in power structures, I don’t > see a problem. I don’t want to automate member participation, I want > participation to be a free choice. Participate or leave it to others. > If you want, you can choose who the others are. If you don’t, well, > that is your prerogative, it is no skin off our teeth. Participate > whenever it suits you, that is what suits us.
-M: “Automate member participation” vs. “participation to be a free choice”.
Did it occur to you that the automation of member participation would be free choice participation?
So if the voter gave the free choice not to select proxies, that the computer would automaticly select underdogs as proxies.> > >L: Yes. But I’m not going to argue the point. If Mark is right, why, of > > > course, all we have to do is to go be selected by an SD2-S based Wiki. > > > But what is going to create that wiki? > > > >-M: Maybe Markus. He has asked me to provide specifications. > >I am still thinking.
>L: So you have nothing. You are just thinking, which is even less than > communicating. Sauce for the goose, Mark.
-M: I did post most of my idea here already, and Markus has a good idea of what I am up to.
Its a Ruby on Rails based Project Management Wiki.
It will be inspired by (and may borrow code from): Pimki, Parlement, Basecamp, Typo, Smartocracy and Instiki.>L: Creating the wiki is still not creating the organization. There are > a number of nearly empty attempts around the internet. I think of > Mikael Nordfors — one of the independent inventors of delegable > proxy, I think — and Nornia.org. I just tried to look at nornia.org > but it did not respond. Last time I looked, Nornia.org was a few > people having made a few comments. The intention was to set it up as > a “delegated voting” organization — he does not use the term > “proxy,” he calls proxies “advisors.”
-M: This is because he is using lemming-DP. Low-order DP is populistic.
High order DP (four or more proxy chains deep) allows for better filtration and peer selection.>L: You can open the doors, but unless people come in, you don’t have a > party. You’ve got some open doors.
-M: It could be marketed as being competitive. People’s general ranks could be posted.
> > >L: Where is one? Show me! > > > >-M: I think that I have provided more specifics for SD2-S than you have > >for specific DP systems.
>L: In other words, what has been provided is theory, which is what I’ve > said I’ve been providing. There is no “input field” for SD2-S, …
-M: I posted this here already: —————————————————————————————————————--
Name ___________________________]
Manditory representitives: [(default), (default)]
Optional additional representitives: _______________________] ____________________________________________________]
Issue X Y Z
Vote:
Yes [ ] No [ ] Deliberate [ ]Optional delegate(s) (will default to representitives if none are selected): ________________________________________]
Decision threshold 60%… (range 50%+1 – 70%, default 60%)
…of a required 20%5 of qualified voters (range (0% – 40%)5, default 20%+5) —————————————————————————
>L: …there is a plan to create one, some day. Yet Mark behaves as if what is > being compared is a clearly established system, and what others are > proposing are mere theories.
-M: Yes, my system is established more clearly than others.
>L: Dysfunctional, again. I do not recall the specifics Mark is talking about.
-M: Then use the search option. I found the input field by searching “input field”.
>L: If this is a coherent theory, more than what has been provided by me, here and on > beyondpolitics.org
- and what is here is generally more developed -> then where is it expressed? Where is the system design?-M: It is here. Search ‘input field’ and ‘constraints’.
[…] > > >L: Because what Mark wants has been tried so many times. Collecting > > > large resources for central control attracts parasites. They will > > > figure out a way to corrupt the system.
> >-M: Will – this pisses me off.
>L: Good. I like pissing off arrogant pretenders. It’s a fault of mine. > I apologize.
-M: You are assuming that I can’t develop an incorruptable filter. You are being disrespectful and presumptuous.
> >L: You are assuming that SD2-S is corruptable simply because it is > >centralized. Yet you don’t understand it, do you?
>L: No. I’m basing this in an inference. It could be wrong. Show me.
-M: OK – inference, which is uncertain.
>L: I do have some ideas about what SD2-S is, but I have not seen sufficient > definition to claim to understand it.
-M: Look at the above input field. The general trustees stay the same, but the specialists change for each issue.
Now imagine the proxy chains to have unlimited depth. This gives a trust network for each issue, as well as a trust network of generalists.> > >L: The safeguard against this: the decentralization of power… > > > >-M: Then it no longer has the benefits of centalization. > >Lets cure or headaches by cutting off our heads.
>L: No, FA/DP does create a centralized structure, but it does it as a > fractal. Smoke that, Mark. If a human brain — which is fractally > organized — is “centralized,” then FA/DP could be said to be > centralized. But there is no “center.” There is a gestalt that > appears, we call it “identity.” Technically, any Buddhist would know > that it does not exist, it is an illusion. But decisions get made > anyway, as if there were a central identity.
-M: We Shaivites do believe in a Self. By parallel, an organization evidences its centrality by making a decision,
no matter how ‘decentralized’ it claims to be.[…] > > >L: Mark, you can be so obtuse. I’d rather be annoying. > > > >-M: Markov chain c-algorithms are easy to understand.
>L: Wow. I must really be an idiot! Presumably they are so easy to > understand that cogent explanation isn’t necessary. Except for idiots > and lemmings.
-M: What don’t you understand? Markov chains? Counting? Random forking? Reiteration?
If you aren’t specific about what you don’t understand, its hard for me to help you.
Here is a primer for SNA:
http://www.orgnet.com/sna.htmlThis won’t tell you about directional Markov algorithms, but it will show you more common c-algorithms.
This will give you a context for understanding the SNA use of Markov algorithms.> >-M: With SD2-S, there would probably be a treasury committee composed > >of high ranking adminstrators, and whose policy has been determined by > >treasury related specialists.
>L: Lots of ways to do it. In FA/DP there is effectively no central treasury. However, caucuses can set up organizations which function in any way they choose. One > of the interesting concepts would be what’s been called the House of > Money. It is called the House because it is like the House of > Representatives, except what is represented is member funding. > Members put funds into an account, and they can “vote” with these > funds. And they can delegate their “vote.” Which would mean that > high-level proxies could quickly marshal and dedicate very > significant funds.
-M: I could see this working. I could also see SD2-S used to employ a similar plan.
>L: However, this is not FA. By definition. It is, however, quite possible as an independent structure, and it could be DP. Structures like this would be the muscles of the body of which FA/DP would be the intelligence.
-M: OK.
> > >L: Can the majority decide to spend the money in the central treasury? > > > >-M: No, but their wishes would be known to the top ranked specialists > >for the given issue. I predict that a reasonable compromise would be made.
>L: So the specialists can spend the money, i.e., a majority of them?
-M: No, the specialists aren’t counted. The specialist vote is determined by PageRank, meaning that anyone recieving ANY proxy votes from ANYONE, has more PageRank/specialist vote than they have popular vote.
SD2-S has seamless continuity between the reps and voters, and this is an anti-entrenchment mechanism.>L: Look, Mark, I think you are making this up as you are going along. I > can understand that.
-M: No. I have done no back-pedalling and my basic concepts are clear.
>L: To some extent, I’m doing that as well. But I’m > not claiming that “the house of money,” for example, is part of the > FA/DP concept. It is one way that it could work out. I think this, > just what you have said: > I predict that a reasonable compromise would be made. > I am not fixing the specific ways in which FA/DP organizations > would function. They will work that out. I’m laying out general > principles, and sometimes, because people ask, showing how these > principles could possibly be applied.
-M: I have general principles – this is SD2. I also have a nailed system, SD2-S.
>L:[…] Proxies are not elected, they are chosen. > Nobody can force you to choose a proxy, a coerced proxy is not a > proxy but a recipient of stolen goods. You don’t have to compete with > anyone in choosing a proxy. It is utterly different from elected > representation.
-M: Cool. With SD2-S, an underdog can even recieve bonus votes.
> > >L: Against the wishes of some of those who contributed it? > > > >-M: No system will appease all the people all the time.
>L: In other words, yes. “Appease” betrays an attitude, possible. He > could have written “please.”
-M: Its is my understanding that democracy imposes whether we like it or not.
And people’s immediate desires may not be what is best for them in the long run.>L: What Mark is describing is standard organizational practice. FAs do > not do this in any significant way, and the concept is a proven one. > It works. Organizational unity is enhanced, which is the purpose of > avoiding this kind of centralization. > > > >L: If so, you have created a disincentive for people to join who might > > > fear that they will be funding something they oppose…. > > > Congratulations. Welcome to same old same old. > > > >-M: Yes, the same old of centralization via democracy. > >There are no other options.
>L: I’ve been describing one with all these words, for all this time. It > seems Mark hasn’t heard any of it. Maybe someone has, now or in the future. > Otherwise, why bother?
-M: You have been describing a parallel system to democracy. By contrast, I have been describing an intended improvement to democracy.
I think the SD2-S is much more interesting than FA/DP, because of its power.
shanti
Mark, Seattle WA USA—~—~————-~—~——~——————~———-~—~——~ You received this message because you are subscribed to the Google Groups “top-politics” group. To post to this group, send email to top-politics@googlegroups.com To unsubscribe from this group, send email to top-politics-unsubscribe@googlegroups.com For more options, visit this group at http://groups.google.com/group/top-politics?hl=en -~—————~——~——~——~———~——~———~—~—-
+1
-
At 06:22 PM 11/3/2006, Mark wrote:
>Lomax wrote: >[…] > >L: Mark expects the world to beat a path to his door. He bought that > > mousetrap myth. Build a better mousetrap and starve, unless you learn > > how to market it or engage someone to market it for you. > > Mark, we are lemmings! Get it? >-M: We are systems designers.
Oh, I am what profession I wannabe?
>I expect there to be few lemmings among us.
We are all lemmings, Mark. That you think you are not — and you do think you are not — shows that you have bought your own exalted image of yourself.
And all of us can break out of leeminghood for times. We can’t do it more than occasionally without the people in white coats taking us away. Been there, done that, by the way.
>-M: My way, everyone is represented, always.
Whether they want to be, or not. You will be represented.
This, by the way, is the status quo. Representatives in the U.S. House are apportioned according to population, not according to votes. The population includes many non-voters, including non-citizens, children, etc.
> > If a member is not represented, so? >-M: Then they aren’t providing useful proxy information. >The voter is the data input for the c-algorithm.
Garbage in, garbage out. If the voter is assigned a proxy, then something provided by the system that assigns the proxy is adding weight to that proxy without any input from the voter.
This is not obvious to Mark because he does not respect the voter in the first place. He’d rather design a system that, instead of amplifying the signal in the voter noise, would substitute another signal provided by the system
We want the choice of a proxy to be by conscious, voluntary choice of the voter. The voter is one of the muscles of society. We want him to function by his choice. Not because the system or some system designer says he must.
We want to end the involuntary alienation of the citizen from government, and the first step is to end that alienation between the citizen from voluntary organizations outside government.
If these organizations are to maximally inclusive, they must not take controversial positions. If they do take such positions, they are merely another competitor in the political marketplace. We want something more than that. We want full participation, and to get it, the organization itself must be universal. This, among other things, means that it must not be a power structure, it must be about communication, voluntary communication.
Adding inputs that don’t come from the citizens just adds, at best, noise, and at worst, it adds bias.
>-M: “Automate member participation” vs. “participation to be a free >choice”. >Did it occur to you that the automation of member participation would >be free choice participation?
No. Because it would not. Automation, perhaps, of the potential for voter participation, perhaps. In other words, the system could make it easy. But if it is too easy, one might wonder if it is inviting unconsidered input. It should be easy. But the system should not make it very easy, as, for example, by presenting the voter with a couple of choices for proxy automatically. It is possible that there are ways such a process could be done without introducing bias.
But doing it by PageRank, if I understand this correctly, will merely amplify the existing bias in the organization, instead of adding the signal from the voter to that bias. There will always be bias, but we want it to be the general bias of the society. Mark might call this “lemmingism,” but that is only a problem when “lemmingism” prevents the consideration of new ideas.
>So if the voter gave the free choice not to select proxies, that the >computer would automaticly select underdogs as proxies.
Those are not proxies. Those are appointed voters. They have been given added power without the excuse that this power was voluntarily and knowingly assigned by the voter. If the system selects “underdogs,” then it is introducing counter-bias of a kind, a signal different from that which, on average, would be presented by the voter. It is, again, noise.
There is no “underdog” problem with FA/DP. Underdogs simply organize themselves and can be as large a force as they can muster. They will be heard. But the system will not balance out underdoggedness such that it becomes equal to topdoggedness. It leaves people as they are, but provides them with means of expression and an assurance that they will be heard.
The automatic assignment of proxies does not set up the relationships of trust that we consider essential to the long-term effects of FA/DP.
If Mark wants to offer automatic assignment tools to organizations, fine. But I think most of us won’t want that crap. He is designing in bells and whistles that his potential customers — at least this one!, and I might have some influence — consider more of a problem than a solution.
He is one of that class of abstract designers of castles in the sky, utopian fantasies. And he refuses to connect with the real world. Which is … lemmings. According to him.
He hates this. Why?
Because they won’t listen to him. If they were lemmings following him, would he mind so much?
Perhaps. And perhaps not.
>Low-order DP is populistic. >High order DP (four or more proxy chains deep) allows for better >filtration and peer selection.
This is an illusory distinction. The only thing that would make Delegable Proxy “low order” as he defines it is a small size of the organization. Large-scale DP will be automatically high order. That is, the average chain length from individual member to top proxy will be more than four. We can easily calculate this.
The Demoex experiment did not understand the implications of delegable proxy, and, while the goal was to create something like universal participation, it did not understand that majority rule within a party excludes universal participation. They did not understand the FA vision, they only had DP. Note that some participants in Demoex have since become informed about this, including some of the founders, I think. We’ll see what the future brings.
We think that in a general purpose political FA, the average direct proxy count will be twenty. If a proxy takes on too many direct proxies, the traffic becomes a burden. Too few is not a problem, really except that it increases the chain length to the center.
If the average proxy count is twenty, the size of the organization is 20 ^ N, where N is the average chain length. (That’s not accurate, I think, because “average” is a linear but we are dealing with an exponential function.) The math to work out the actual distribution is more complex than I care to work out at this time.
If, somehow, we had an actual direct proxy count of twenty, then with N = 4, we’d have 160,000 members. Now, what in the world does this have to do with “better filtration and peer selection?”
It simply scales the organization. The hierarchy created spontaneously is a fractal that is self-similar regardless of scale. The rapid information transfer across the network gets noisy, a bit, as the chain length increases, but these are intelligent filters and links. They can and will reach from the top to the bottom to validate communications, this is not the telephone game, except for rapid communication. That is, there will be accuracy distortion in rapid transmission, where messages are passed on without detailed consideration, but perhaps in the sense that there is an emergency. But this will not be how it ordinarily functions. Rather, proxy filtering opens up a channel from the bottom to the top. There is no distortion, ultimately.
There must be relationships and expectations of trust for this to work. It is not something that can well be automated. Recording proxies, analyzing proxies, notifying members that loops or dead ends prevented them from being represented in a poll, that can be automated.
But my point is that it does not need to be. It can all be done manually, with little difficulty. Demoex made a big mistake, which was in thinking of DP as a computer thing. It’s not. It could have been done hundreds of years ago. Or more. But the communication involved on a large scale gets faster and a bit easier.
> >L: You can open the doors, but unless people come in, you don’t have a > > party. You’ve got some open doors. >-M: It could be marketed as being competitive.
Marketed by whom? By you?
>People’s general ranks could be posted. > > > >L: Where is one? Show me! > > > > > >-M: I think that I have provided more specifics for SD2-S than you have > > >for specific DP systems.
We have very small systems in place. Because they are Free Associations, with specific rules as determined by the members, the specification is so simple that Mark doesn’t even recognize it as such.
(1) Define an organizational purpose. (2) Set up a proxy list. This is just a list, it is not an automated system. (3) Invite members to join and participate. (4) According to standard process (I recommend Robert’s Rules as adapted for on-line meetings), conduct polls. A poll result, again, creates another list. Voter ID, vote.
that’s it. The rest is what the members do voluntarily. If they need more advanced tools, they will find, build, or buy them. If Mark has tools ready, I’m sure they would be considered. But, I can tell you, if they research his writings and find all this “lemming” nonsense, they will turn away, perhaps needlessly.
Mark has understood a great deal, certainly about DP. He understands some dimensions of the problem, and this understanding is unusual. But he is also a lemming himself, and his lemmingness keeps him from seeing beyond certain limitations. He can’t see how an organization that is, in itself, powerless, could radically alter the balance of power.
He does not see the power of advice and information. Neither of these has power in itself.
He also does not see the danger of centralization. His opinions about this, about power in general, are actually the “common wisdom.” His dislike for direct democracy is a very old idea, including his belief that most people are merely sheep. Or lemmings, his preferred term.
People are lemmings, yes. It’s a protection and a strength. But it does introduce certain problems, and we believe that the kind of communication that FA/DP should foster will solve those problems.
> > Mark behaves as if what is > > being compared is a clearly established system, and what others are > > proposing are mere theories. >-M: Yes, my system is established more clearly than others.
No. It is not established at all. It is a design. A design is not “established” until it is in use. One design may be more practical, though less completely specified, than another which has been even completely specified.
A completely specified error is, what? Superior to a design approach that is, necessarily, open and therefore not completely specified?
> >L: Dysfunctional, again. I do not recall the specifics Mark is > talking about. >-M: Then use the search option. I found the input field by searching >"input field".
Whatever time I’m spending here is essentially stolen time. I’m already spending too much. Do not expect it to continue.
>-M: You are assuming that I can’t develop an incorruptable filter. >You are being disrespectful and presumptuous.
Yes. I am assuming that, because I consider it an extremely difficult problem. I don’t think Mark has solved it, at all. But if he is correct, surely his system will gain notice from the millions of people who are searching for one. Or are there millions?
Probably not. There may be three or four. I’ve been one of them. I abandoned the search, because I don’t think it is possible.
FA/DP is not corruptible, or, more accurately, it would be so expensive to corrupt that we would just call it “compensation” instead of bribery. You pretty much have to corrupt a majority, and it is quite difficult to do that without letting the cat out of the bag, so, in practice, you have to offer compensation to everyone.
FA/DP is not a power structure, so corruption in it would amount to proxies offering bad advice. It would amount to low-level proxies offering bad advice. This is because top-level proxies, where information management power is concentrated, cannot easily bypass their clients to reach the end of the chains, the members. They may be able to reach down a few levels, depends on system design, but they will not be able to reach into the direct relationships of trust at a low level. Where there is no compensation and, in a political system, we are probably talking about neighbors and personal friends. How much would you have to pay me to lie to my friends?
And, indeed, if you paid me, how would you know that I had actually gone ahead and lied? My communication with them is private and privileged, if I want it to be. Proxies and clients can and will communicate independently of the central network, which exists only a sort of clearing-house.
Mark is assuming that he can develop a centralized but incorruptible system. It has not been done, it has never been proven. My assumption, that he cannot, is highly likely to be correct. It is not impossible that he could succeed, but I would not bet the time of day on it.
The value of this conversation to me is that it exposes certain ideas about FA/DP that might not otherwise be made explicit. This is what I’m doing most of the time, at this stage of our process.
> > >L: You are assuming that SD2-S is corruptable simply because it is > > >centralized. Yet you don’t understand it, do you? > >L: No. I’m basing this in an inference. It could be wrong. Show me. >-M: OK – inference, which is uncertain.
Yes. Now, show me a deductive proof that SD2-S is uncorruptible. In the absence of that, induction is all we have. From what evidence is Mark inducing that it is incorruptible? His own ideas about it? Or does he have a rigorous proof?
If he does, from what he has written before, I’d guess that it is written in such technical language that I am not going to review it myself. Rather, when the time is right, I’ll sic some mathematicians or other specialists on it.
Otherwise, what is more likely, is that he has, in his mind, ideas about how to prevent corruption, and we will consider those ideas if they seem reasonable.
>L: I do have some ideas about what SD2-S is, but I have not seen sufficient> > definition to claim to understand it. >-M: Look at the above input field.
I did. First thing I see is “manditory representatives.” Mispelled, by the way, Mark should learn to use spellcheck. What do I do if I am undertain about joining this organization and I see that?
I go somewhere else. I don’t like being subject to mandates, period.
>The general trustees stay the same, but the specialists change for each >issue.
This is utterly impractical. The “lemmings” cannot manage complex issues outside their individual areas of focus, they simply do not have time. Even to name specialists.
We do have a mechanism for “issue proxies,” which is that independent FA/DP organizations will exist, each one devoted to an issue. There may be one or more metaorganizations to which everyone would belong, and general proxies from these organizations would manage membership, effectively, in suborganizations where the member does not wish to directly participate.
We wish to keep the proxy system extremely simple, and this requires single proxy, with default proxies, those active in the absence of the single proxy, coming from proxy assignments of the absent proxy, thus concentrated. Encouraging the designation of alternate proxies destroys this simplicity.
*But members can do it. Its a Free Assocation, you can do whatever you like. But getting the organization to do the very much more complex analysis involved, that would be another thing. I don’t think it will happen.
The most we see is what was suggested for Metaparty. (metaparty.beyondpolitics.org) There are Forums there for the consideration of various issues. Any participant in a forum can set up a special proxy list. This list will override the standard proxy for that forum, assuming that analysts want to use it. But this is really a matter of an individual “meeting” setting up its own rules. It has no effect on other meetings.
Issues will be constantly changing. We don’t want to encourage issue proxies because it dilutes the decision to trust the single proxy. We want that decision to be made carefully. We want it to be general trust, not very narrow.
However, FA meetings can and will consult specialists, and there will be whole organizations of specialists, perhaps with membership requirements such as credentials. These latter organizations can be FAs within their membership definition. (Generally FAs allow self-definition of membership qualification, and these organizations might be the same but might require statement of credentials, if any, and might provide for investigation of the statements. You can join and claim you are a qualified physician, for example, but analysts interested in consensus among physicians are going to discount your claim if you have not provided verifiable evidence. You don’t have to provide that evidence, you can do what you like. But we don’t have to give any particular credence to your votes.
Once again, that this is not a power structure allows “decisions” to be made without needed to formally validate membership qualifications. A minority caucus is totally free to go ahead and act without the permission of the majority. Neither one may act as if they represent the FA. They act on their own.
Again, some central tools may be provided, perhaps a means to expand votes using various databases as input. That is, one might determine such things as direct vote results, generally expanded vote results, validated member results (from open or private lists of validated members), etc.
We are unconcerned about member fraud, sock puppets, more than considering it a minor nuisance. It would make things a little more complicated, that’s all. But, over time, the sock puppets will stand exposed. They can’t show up at a face-to-face meeting, for example, they can’t contribute funds that were not available to the real person behind them, they can’t muster votes in real elections. If the central organization decides to validate memberships in some way (it can do this by majority vote; and a caucus suspecting vote fraud may also be able to influence a trustee to allow it even if the majority opposes, trustees are independent. Admin consists of trustees), it can do it, it can send direct emails to members. It won’t do this more than very rarely. Harassing members is a good way to destroy an FA…. or to split it, which is much less harmful.
If a minority suspects fraud is creating a false majority, it can simply form its own FA. The false majority ends up holding a handful of sand. And if it was not false, it was real, so what? The minority, in leaving, will leave behind proxies collectively representing them. There is still an overarching FA, the original one.
Meetings multiply and remain connected.
Unspecified? Sure. In detail. But the vision is quite clear, and I know this because others have been “infected” by it. They can write about this topic and I look at it and say, “That’s what I would have said.” They are describing the same vision.
Mark has got a few aspects of this vision, and in particular, much of the reasoning behind DP. But he does not know how to get from here to there, yet he is trying to design a system that will be used when we are “there.” I.e., when power structures are being operated with DP methods.
What I’m saying is that this is way premature. FA/DP will gather the intelligence to design — or accept — advanced social decision-making systems, and will coordinate the power necessary to implement them. In doing this, it must be free of bias as to the outcome. Bias is the enemy of intelligence. Bias an intelligent process and you bias the result away from optimum.
>Now imagine the proxy chains to have unlimited depth.
We imagine proxy chains to have unlimited depth, it goes without saying in delegable proxy. If proxies are delegable, there is no limit to the delegation, until it loops back. Theoretically, if everyone names a proxy, there is a limit to the chain depth, it is the human population. But, of course, information would move very, very slowly across that “network.”
>This gives a trust network for each issue, as well as a trust >network of generalists. > > > >L: The safeguard against this: the decentralization of power… > > > > > >-M: Then it no longer has the benefits of centalization. > > >Lets cure or headaches by cutting off our heads. > >L: No, FA/DP does create a centralized structure, but it does it as a > > fractal. Smoke that, Mark. If a human brain — which is fractally > > organized — is “centralized,” then FA/DP could be said to be > > centralized. But there is no “center.” There is a gestalt that > > appears, we call it “identity.” Technically, any Buddhist would know > > that it does not exist, it is an illusion. But decisions get made > > anyway, as if there were a central identity. >-M: We Shaivites do believe in a Self. >By parallel, an organization evidences its centrality by making a >decision, >no matter how ‘decentralized’ it claims to be. >[…] > > > >L: Mark, you can be so obtuse. I’d rather be annoying. > > > > > >-M: Markov chain c-algorithms are easy to understand. > >L: Wow. I must really be an idiot! Presumably they are so easy to > > understand that cogent explanation isn’t necessary. Except for idiots > > and lemmings. >-M: What don’t you understand? Markov chains? >Counting? Random forking? Reiteration? >If you aren’t specific about what you don’t understand, >its hard for me to help you. >Here is a primer for SNA: >http://www.orgnet.com/sna.html >This won’t tell you about directional Markov algorithms, >but it will show you more common c-algorithms. >This will give you a context for understanding the SNA use of Markov >algorithms. > > >-M: With SD2-S, there would probably be a treasury committee composed > > >of high ranking adminstrators, and whose policy has been determined by > > >treasury related specialists. > >L: Lots of ways to do it. In FA/DP there is effectively no > central treasury. However, caucuses can set up organizations which > function in any way they choose. One > > of the interesting concepts would be what’s been called the House of > > Money. It is called the House because it is like the House of > > Representatives, except what is represented is member funding. > > Members put funds into an account, and they can “vote” with these > > funds. And they can delegate their “vote.” Which would mean that > > high-level proxies could quickly marshal and dedicate very > > significant funds. >-M: I could see this working. >I could also see SD2-S used to employ a similar plan. > >L: However, this is not FA. By definition. It is, however, quite > possible as an independent structure, and it could be DP. > Structures like this would be the muscles of the body of which > FA/DP would be the intelligence. >-M: OK. > > > >L: Can the majority decide to spend the money in the central treasury? > > > > > >-M: No, but their wishes would be known to the top ranked specialists > > >for the given issue. I predict that a reasonable compromise > would be made. > >L: So the specialists can spend the money, i.e., a majority of them? >-M: No, the specialists aren’t counted. >The specialist vote is determined by PageRank, meaning that anyone >recieving ANY proxy votes from ANYONE, has more PageRank/specialist >vote than they have popular vote. >SD2-S has seamless continuity between the reps and voters, and this >is an anti-entrenchment mechanism. > >L: Look, Mark, I think you are making this up as you are going along. I > > can understand that. >-M: No. I have done no back-pedalling and my basic concepts are clear. > >L: To some extent, I’m doing that as well. But I’m > > not claiming that “the house of money,” for example, is part of the > > FA/DP concept. It is one way that it could work out. I think this, > > just what you have said: > > I predict that a reasonable compromise would be made. > > I am not fixing the specific ways in which FA/DP organizations > > would function. They will work that out. I’m laying out general > > principles, and sometimes, because people ask, showing how these > > principles could possibly be applied. >-M: I have general principles – this is SD2. >I also have a nailed system, SD2-S. > >L:[…] Proxies are not elected, they are chosen. > > Nobody can force you to choose a proxy, a coerced proxy is not a > > proxy but a recipient of stolen goods. You don’t have to compete with > > anyone in choosing a proxy. It is utterly different from elected > > representation. >-M: Cool. With SD2-S, an underdog can even recieve bonus votes. > > > >L: Against the wishes of some of those who contributed it? > > > > > >-M: No system will appease all the people all the time. > >L: In other words, yes. “Appease” betrays an attitude, possible. He > > could have written “please.” >-M: Its is my understanding that democracy imposes whether we like it >or not. >And people’s immediate desires may not be what is best for them in the >long run. > >L: What Mark is describing is standard organizational practice. FAs do > > not do this in any significant way, and the concept is a proven one. > > It works. Organizational unity is enhanced, which is the purpose of > > avoiding this kind of centralization. > > > > > >L: If so, you have created a disincentive for people to join who might > > > > fear that they will be funding something they oppose…. > > > > Congratulations. Welcome to same old same old. > > > > > >-M: Yes, the same old of centralization via democracy. > > >There are no other options. > >L: I’ve been describing one with all these words, for all this time. It > > seems Mark hasn’t heard any of it. Maybe someone has, now or in the future. > > Otherwise, why bother? >-M: You have been describing a parallel system to democracy. >By contrast, I have been describing an intended improvement to >democracy. >I think the SD2-S is much more interesting than FA/DP, because of its >power. >shanti >Mark, Seattle WA USA >
—~—~————-~—~——~——————~———-~—~——~ You received this message because you are subscribed to the Google Groups “top-politics” group. To post to this group, send email to top-politics@googlegroups.com To unsubscribe from this group, send email to top-politics-unsubscribe@googlegroups.com For more options, visit this group at http://groups.google.com/group/top-politics?hl=en -~—————~——~——~——~———~——~———~—~—-
+1
-
Lomax wrote: >Mark wrote: > >Lomax wrote: > >[…] > > >L: Mark expects the world to beat a path to his door. He bought that > > > mousetrap myth. Build a better mousetrap and starve, unless you learn > > > how to market it or engage someone to market it for you. > > > Mark, we are lemmings! Get it? > >-M: We are systems designers.
>L: Oh, I am what profession I wannabe?
-M: Its a hobby for now.
> >I expect there to be few lemmings among us.
>L: We are all lemmings, Mark. That you think you are not — and you do > think you are not — shows that you have bought your own exalted > image of yourself.
-M: I do have an exalted image of myself. I rock. How am I a lemming?
>L: And all of us can break out of leeminghood for times. We can’t do it > more than occasionally without the people in white coats taking us > away. Been there, done that, by the way.
-M: I hope that you were the one with the white coat. I want to be permenently out of lemminghood.
> >-M: My way, everyone is represented, always.
>L: Whether they want to be, or not. You will be represented. > This, by the way, is the status quo. Representatives in the U.S. > House are apportioned according to population, not according to > votes. The population includes many non-voters, including > non-citizens, children, etc.
-M: That isn’t necissarily representing me. By contrast, with SD2-S, all voting power(unless looped), becomes a quantified part of the trust network. It doesn’t get thrown out just because its not part of some majority.
> > > If a member is not represented, so? > >-M: Then they aren’t providing useful proxy information. > >The voter is the data input for the c-algorithm.
>L: Garbage in, garbage out. If the voter is assigned a proxy, then > something provided by the system that assigns the proxy is adding > weight to that proxy without any input from the voter.
-M: In that case, its being added by other voters. With your option, where no proxy is provided, its still other voters’ voice being provided instead(that is all that is counted). The only difference is that my approach attempts to smooth the rank distribution curve by giving bonus votes to underdogs. This makes the rank structure less clustered and entrenched.
So SD2-S actually does something with the voter’s proxy power instead of ignoring it.>L: This is not obvious to Mark because he does not respect the voter in > the first place.
-M: You are the one who wants to ignore the voter’s proxy power instead of doing something constructive with it.
>L: He’d rather design a system that, instead of amplifying the signal in the voter noise, would substitute another signal provided by the system
-M: No, I am trying to amplify the signal in the voter noise, and my system cannot create substitute signals. It sees only the rank distribution curve, and looks for anomolous gaps to fill. This also makes the ranks more differentiated, and therefore the data more useful.
>L: We want the choice of a proxy to be by conscious, voluntary choice > of the voter. The voter is one of the muscles of society. We want him > to function by his choice. Not because the system or some system > designer says he must.
-M: Yes, with SD2-S the voter is given the choice of real people or defaults. With your preference, the voter is given the choice of one person or being ignored
>L: This, among other things, means that it must not be a power structure, it must be about communication, voluntary communication. Adding inputs that don’t come from the citizens just adds, at best, noise, and at worst, it adds bias.
-M: “inputs that don’t come from the citizens” - Again, you are not describing SD2-S.
> >-M: “Automate member participation” vs. “participation to be a free > >choice”. Did it occur to you that the automation of member participation would > >be free choice participation?
>L: No. Because it would not. Automation, perhaps, of the potential for > voter participation, perhaps. In other words, the system could make > it easy. But if it is too easy, one might wonder if it is inviting > unconsidered input. It should be easy. But the system should not make > it very easy, as, for example, by presenting the voter with a > couple of choices for proxy automatically. It is possible that there > are ways such a process could be done without introducing bias.
-M: Your example here would be ‘too easy’ and would add the bias that you describe.
SD2-S would not add recommendations unless they are calculated to unentrench the rank structures.>L: But doing it by PageRank, if I understand this correctly, will merely > amplify the existing bias in the organization, instead of adding the > signal from the voter to that bias.
-
-
04/11/06 00:22