Wiki Stone Society
Wiki Stone SocietyWikiWikiWeb has grown and thrived along with the web itself. But we all wonder, are there limits to this growth?
PeterMerel describes the StoneSociety as a possible mechanism for administering an even larger wiki without resorting to concentration of control. In this page I would like to explore further exactly how a next generation wiki could exploit the ideas presented in his paper. Contributors to this page are encouraged to read and reflect on that work first. The following attempts to use terms exactly as defined therein. -- WardCunningham
What is this
A �stone� society is a meritocratic system to manage shared resources.
The Society's scope and purpose are defined by a Charter. Within a society tokens (stones) are allocated to authenticated people who take responsability for a certain duty. Choices are made by means of auctions with those tokens. So you get more and less important contributions from important or less important contributors.
Wiki in Stone Terms
A wiki web is a resource that a society chooses to administer. The charter for such a society defines the purpose and limits to the wiki. The resources of a wiki (both physical and intellectual) can be partitioned and repartitioned as appropriate at any given instant in time.
A society administers a wiki by investing power in officers who accept commissions and therefore responsibility for some partition of a wiki's resources. Power is counted in stones, a novel surrogate for money that is made possible by and exploits unique properties of electronic communication.
Power, and with it all important decision making, never actually leaves the hands of the members of the society. Commissions are only for a term, and even this can be revoked through a group decision making process called an auction.
Mechanics
Each page of the Wiki would be unmoderated until an Officer of its Society created an Auction to moderate it. An Auction would be desirable not in order to stifle debate, but to control which content gets associated with the page's WikiName. I'm thinking then that each Auction would have to be associated with two main pages; one the new moderated page, and one the old unmoderated page where unfettered debate could proceed.
These two pages would be mutually linked, and the moderated one would also include links to pages that represented each of the Preferences available for its Auction. These, I think, would have to be unalterable once proposed. Stone Ante conditions could be used to discourage the creation of many frivolous "noise" Preference pages, and the Stone "donkey" preference would represent the option that the WikiName be unmoderated.
Participants could then register various levels of interest in moderated wiki pages. They could subscribe to all changes, to Preferences that achieved bids that threatened to make them become Quanta (good for keeping up with bidding contests), or just to actual changes of Quantum.
The key effect of all this is that, once the first Term of the Auction has passed, the WikiName would no longer link to the unmoderated page, but to the moderated page, whose content would become whatever content is associated with the Quantum.
Note that the proposer of a Preference does not gain editorial control if his Preference becomes Quantum. To do that would be to delegate authority again, which is just what this system sets out to avoid. Any person at all who wants any change made would have to go though the process of bidding it past the present Quantum - which would generally require consensus from other interested Officers.
-- PeterMerel.
How would wiki be more valuable after the imposition of all this overhead? -- KentBeck
Good question.
-- PeterMerel.
I am trying hard to work with you on this, still not succeeding. Part of my difficulty is seeing a phrase like "Examples of Wiki Problems Solved" which solves non-problems by adding work. I did not manage to see how the housekeeping task vanished when you made two pages out of one and held auctions.
Wiki is currently based on the idea that people really can be polite and well mannered. It is true that such an idea falls apart when the group loses its sense of community. So I am willing to go along for the ride.
I read the StoneSociety and perhaps should read it together with a friend to sort it out, but two crucial things I couldn't figure out... how are stones manufactured, and who gets to become an Officer? Both those struck me as rather authoritarian, and I did not see the answer. Thanks.
An idea like this, with no working model - not even some proof-of-concept game to work through - not even a diagram - I'm not surprised it's hard to see. I'm working on getting these down in bits, and I'll let folk here know when I do. I broached the StoneSociety idea on the Wiki in passing and was surprised (and delighted) to see WardCunningham pick up on it with this page. But I think it should get easier to explain as I get more material online - I plead your indulgence.
On "wiki problems solved", I think that's Ward's text - it wasn't signed, but it wasn't mine. [I wrote the heading as much as a question as a statement. Then I took some guesses. -- Ward] Ward's idea appears to be that you'd use Auctions to determine who is moderator - who controls a page as a Resource. I replied with the notion that with Auctions you needn't even have a moderator; that control of the Resource can be better constrained than that.
Then, as Alistair says, people can be polite and well-mannered. The aim here is to find a way to scale the properties of humans getting along by manners up to a large scale - to scale what the social psychologists call "syncracy", one of my favourite words.
On the specific questions:
As to authoritarianism, I think Stone Societies should end up pretty anti-authoritarian so long as there are no large distinctions between the numbers of Stones that come with various Commissions; the numbers of Stones in each Commission is a matter for definition by Charter and by the market, not an inherent property of the system.
-- PeterMerel
Stone Societies seem to predicate all kinds of configuration and limits on what's called a Charter. What would a WikiStoneCharter look like?
RonJeffries writes:
And PeterMerel replies:
Seems to me this notion builds haves and have-nots. Of course we would be haves.
There's nothing inherent in the StoneSociety to build haves and have-nots. It is perfectly feasible to base a StoneSociety on equal-sized Commissions, and for such a StoneSociety to require no qualification beyond voluntary acceptance for the issuing of these Commissions. That'd be purest democracy. The notion is just that you structure your society as convenient to the purpose of its members. What would be convenient here? That's the subject of WikiStoneCharter. The StoneSociety as a mechanism imposes no such structure of itself.
In my experience, systems work best if they, by their nature, encourage "good" behavior and frustrate "bad".
Very much agreed. There's a fairly extensive analysis in these terms on the Stone website you can get to via StoneSociety.
Perhaps a few system features and social pressure can help here more than an elaborate new form of society. I am thinking of things like better backups of pages so they can't get destroyed; easy features for splitting a page into the gelled part and the random part.
[Please forgive me for butting in here. Check my drivel on PerpetualNow for a little discussion on this. And now, back to our regularly scheduled discussion. -- MartySchrader]
The former is already available in the CvWiki patch, which integrates Wiki with CVS. The latter requires a mechanism for doing the gelling, which is what's being proposed here. Can you think of another one?
I do share the concern: in a recent email to Ward, I asked whether improving wiki's responsiveness might actually be counterproductive by letting it go out of control.
I should say that the CvWiki stuff seems to perform well with scale, but I understand Ward has also got 2 or 3 different improved versions almost ready to roll. Just when they roll out as c2.com/cgi/wiki is a good question.
Wiki's strength is in its freedom, I suggest.
It's difficult to see how WikiStoneSociety would be less free - can you explain why it might be?
WikiStoneSociety would be something else. It might be a good thing, and we should try one - but it wouldn't be wiki.
What difference seems most significant to you?
DaveHarris writes,
To me the significant feature of a StoneSociety is that it requires you to rate different auctions against each other. You only have a limited number of stones, so you have to choose which auctions to vote them in. (If you have more stones then auctions, it is about the weight of your vote rather than whether to vote at all - same difference.)
Can you explain why this feature would suite Wiki pages? Why should voting on ThisPage reduce the weight of my vote on ThatPage?
Because this way you'll exercise some judgement about the merits of each page. The notion is not just to poll people's preferences, but to quantify them. With Stones, if I'm violently opposed to a measure, or violently supportive of it, I can fling my entire franchise that way. If I'm just lukewarm, I can express that. If I'm conflicted, or if I have another way, I can plump for those too - or all of the above. It's a free market.
Presuming we allow fractional Stones, or we distribute sufficient for you to divvy them up evenly per Auction, you can still vote them as if you have no greater preference for ThisAuction? than ThatAuction?. But Stones gives you the other options too because they're a good way to express a rational opinion.
Do you agree that this comparative rating of pages is the essence of the StoneSociety approach?
It's one of the essences. Another way to regard this aspect is that the StoneSociety represents Resource-sharing as a marketable commodity. The Stone mechanism, however, goes further than this. By recycling Stones at the end of each Term, a StoneSociety facilitates the continual reassessment of Resource control. By permitting bids *against* as well as for particular Auction alternatives (here, contents of pages), the StoneSociety permits preemption of Resource control.
These three aspects are, so far as I'm aware, not provided by any other single system. Under statism a government is able to effect the third of them; under capitalism a market is able to effect the second of them; under socialism, cooperatives are able to effect the first of them. The essence of the StoneSociety is to consistently and equitably provide all three.
I think that it would be nice if Wiki had a variety of access policies, naturally enforced by using computers and networks and encryption and such like, and I see StoneSociety as just one possible policy. I aim for SpikeWiki to develop a more general system of which a WikiStoneSociety would be just one incarnation, one kind of group charter. I personally don't see it as being the most important or interesting. -- DaveHarris
I can see how this can be done, but I'm also curious about the alternatives you envision. Can you be more explicit Dave?
Well, for example using one-man-one-vote. Track the votes, allow voters to change their opinions at any time - there's no need for fixed auction periods. Everything becomes simpler and freer. I feel that a lot of complexity of StoneSociety is trying to work around problems caused by the conception of stones. -- DaveHarris
RaySchneider writes,
I'm not sure that I read all the discussion of StoneSociety and Wiki Stone Society -- but the Wiki system here is too open in one sense ... the freedom to change other author's patterns is certainly a temptation to malicious behavior. Indeed; that it hasn't happened here so far is a testament both to the power of etiquette and the quality of the participants.
One might be able to deal with that through a hierarchical structuring:
I like this scheme up to a point, but I think it can give rise to unresolvable polarization. If your community splits on some issue, say an ethical difference, then this scheme gives us no way out. People will politic and things will get both noisy and nasty; an editorship will soon rise and then we're back to representatives.
It's better than nothing, but WikiStoneSociety seems like a sturdier alternative.
RaySchneider's discussion of future wiki requirements moved to WikiFuture
RonJeffries continues his line of questioning:
Wiki's strength is in its freedom, I suggest [...] WikiStoneSociety would be something else. It might be a good thing, and we should try one - but it wouldn't be wiki.
I'm still puzzling over the qualitative difference you see. Would a WikiStoneSociety be less wiki than some of the existing and proposed WikiClone(s)?
What difference seems most significant to you?
I may not understand WikiStoneSociety, he said understatedly. Today I've not been elected, don't have a commission, don't have any stones, yet I can type here and (attempt to) add value. In a WikiStoneSociety (as I understand it) I can't, unless some group process authorizes me. Or am I just confused? -- RonJeffries
Yes and No :-) On some wikis, such as OrgPatterns, you do presently have to be authorized in order to add value. Others, notably this and JosWiki, don't require anything but connectivity. WikiStoneSociety will cater to both kinds; it's a matter for WikiStoneCharter(s) to decide what the Commission conditions are.
But the StoneSociety mechanism is actually more flexible than this, supporting networks of interoperating societies each with different conditions on franchise. Just how such networks will best be configured - whether they'll look like trees, DAGs, layers, PipesAndFilters or what - is an open question. In general, the interconnection of Wikis is something that hasn't been given enough attention. See InterWiki for a brief treatment of such issues.
(Moved from ReproductiveEthics)
Correct me if I'm wrong: under WikiStoneSociety, a random individual cannot edit a randomly chosen page, as we can on wiki. Respondez! -- RonJeffries
The notion is that anyone can edit any page until contention arises. The contenders can then place the content of the page under trial-by-market (a Stone Auction) to resolve things. One of the available options in this Auction would be the donkey preference - to make the page unrestricted again. And anyone could still make any edit to an uncontrolled page where discussion relevant to an auction would be placed.
The notion is not to restrict edits, but to restrict the association of content with WikiName(s). This is more restrictive than the current laissez-faire regime, but it's less restrictive than, say, the editorial review presently applied to publication of software patterns.
-- PeterMerel
At last, I get it. This proves that if you make a thing simple enough, even I can get it. I think I like the idea better now. Thanks! -- RonJeffries
However, I'm not as smart as Ron. This stone society stuff has always been very difficult for me to follow, in part because I'm not too bright, and in part because it's presented in very abstract terms. The above discussion helps a bit, but it doesn't go far enough for me. I'd like the discussion of a supposed WikiStoneSociety to be much more concrete. Not necessarily definitive or predictive, but concrete. So, let's see WhatMightHappenInaStoneWiki.
It is really interesting to read about the stone society, having wandered over from http://everything.slashdot.org. Everything just came up, and the implementation is in a high degree of flux as features are nailed down.
The implementation at Everything requires user to log-in and labels content they create. It is very limiting because only the original author may edit the work. An alternative definition may be posted and voting selects which one stays, theoretically weeding out "bad pages". The problems observed from time spent on http://jos.org was not of bad pages, but of bad changes to pages. Changing somebody else's page is a crucial feature of Wiki (as near as I can tell). The only mechanism I can think of for prevention over "bad" changes without the full freedom of wiki, would be to give the original author veto privileges.
-- PehrAnderson?
Well, you could try the CvWiki/AtisWiki WayBackMode idea. The trouble with it is that "bad" changes still live some place, even when the current version is cool. Not much of a drawback, but it does seem to go against WhyWikiWorks.
[You could also introduce authorship vs editing. This goes a long way toward intellectual property preservation. At altima.org, they have both. An author may only add or remove themselves. Multiple authors are allowed and retained even if an editor attempts to strip them from the page.] -- BobRacko
First, much of the thrust here seems to focus on the eventual collapse of this wiki under the weight of its participants. My question is this: is it really the case that this wiki will grow so large that the participants will destroy it? I know, some will claim that it's already happened. I can read that in between the MustEverythingOnThisWikiBeExtreme lines. Certainly, it's true that this wiki isn't just a pattern catalog anymore (if it ever was). But that could (and seemingly did) come about as much from a demographic shift as from a growth spurt.
I don't know. WikiStoneSociety still seems like an interesting idea to me, but I certainly see no need for it here yet. Then again, my focus for StoneSociety stuff isn't particularly Wiki-oriented these days. Also, though perhaps this is only my perception, the C2 wiki doesn't seem to be growing in population very fast. It may be just a little busier than it was when I first came here maybe 18 months ago, but not a lot. Certainly it's experiencing nothing like exponential growth.
Second, communities come and they go. Usenet used to be a hopping place, for sure, and now its a noisy one. But I remain one of those persons who welcomes the unwashed masses into my playground. Sure, the arrival of the barbarians brought MakeMoneyFast and ComeToBettysDungeon?, but it also brought nationwide cheap (almost free) connectivity, a rather extraordinary value. And though the original Usenet hardcore isn't on Usenet any more, it's not like they all died. They just moved on. One place you might currently find them is right here. I can't speak to where they'll be five years from now if this wiki sells like Usenet and WWW did. But they'll be somewhere, and I imagine I'll be hanging out with them.
Sure.
Third, why not just make the place a gated community? Which of the problems that might be solved by a WikiStoneSociety wouldn't also be solved by making any given wiki a private club?
[This has been tried at http://mozilla.devel.org with limited success. There are public pages and private ones. Anyone may register for free and the editing password is emailed to you (ensuring minimal accountability). It was tried again at http://www.altima.org where wiki is used throughout the site (though you would never notice if you weren't sensitive to the word "wiki" appearing in content)] -- BobRacko
For me, it's just like you said in your second paragraph. There's a real difference between private clubs and open forums, and I much prefer the latter. Plus there's always a challenge in making things scale, and challenges are fun.
Fourth and finally, what are we really risking if this wiki collapses? I live in an intentional community whose members have invested hundreds of thousands of dollars, built their own dream homes, and put at deep risk practically everything typical Americans hold dear. I understand their need to minimize risk, though I usually disagree with the extent of that need as well as the efficacy of the risk-limiting agreements. But what do we have to lose here? A few thousand wiki pages, some of which are outstanding, some of which are crap, and a good many of which are little more than pre-publication review copies of papers PLoPD would reject. I'm not sure, but I think that all the best content of this wiki has grown precisely from its remarkably anarchic nature. I worry that efforts to harness this horse to keep it from running wild will ultimately keep it from running at all. -- MichaelHill
What would we lose if we lost the wikiweb? Well, for a start, a community. Good communities are hard to find. Plus we'd lose the interconnections between the pages, which imho are more important than the isolated content. I wouldn't read wiki if it were a book - all those little pages would be as annoying as hell - but the SmallWorld links here make the content much more valuable. And we'd lose the group-mind - the well trod pathways of discourse and feedback that make this a good place to float an idea and watch it get shot at.
But in reality we wouldn't lose anything. Wiki would just reincarnate somewhere else. The discontinuity wouldn't be pleasant - just today I went looking for the InterWiki stuff that was on TomStambaugh's apparently defunct MuWebWeb and was annoyed not to find it - but wiki would self-assemble again. It is immortal, after all. -- PeterMerel
Rather than implement a system of privileged moderators, would it not be possible to invoke some kind of collective scoring of individual contributions? I know it's outside the technology of the current interface, but imagine a system where anyone could vote on each individual contribution to a WikiName. A positive vote makes that particular contribution somehow more permanent - it stands out more, gets brighter. A negative vote dims it down - it gets grayer or smaller, or becomes an icon you click to read the text. With enough negative votes, it gets swept into an other stuff related to this page bin, archived and still accessible, but out of the flow. Gradually, each page takes on the personality of its readers.
A potential downside is that page content might tend to become bland over time, as more extreme positions got voted down, but I'm not convinced that would happen. -- DaveThomas
I believe the system of privileged moderators is a non-starter, but I'm concerned that voting on each edit is too thread-oriented for wiki. A single edit can range all over a page tweaking and extending all kinds of bits and pieces. The notion of binding WikiName to frozen content isn't much better, though, because the debate will simply move to an unfrozen page with a paraphrased name, and links will be morphed to that. The only meaningful way to enact a WikiStoneSociety then must entail some hierarchical view of content, perhaps something like that found in ZwiKi ... or maybe something stronger still. Very un-wiki-like, in other words. -- PeterMerel, 2002.
I missed an important factor in the discussion for creating such a society: CollectiveLove.
Cited from above:
Wiki's strength is in its freedom, I suggest. / It's difficult to see how WikiStoneSociety would be less free - can you explain why it might be? / WikiStoneSociety would be something else. It might be a good thing, and we should try one - but it wouldn't be wiki. / What difference seems most significant to you?
To use Wiki with any regulations, i.e. StoneSociety, would change the character and the spirit of it completely! I want to warn of the temptation of taking over any control or give up trust and faith in the holy spirit of the world. The main attraction is the miracle WhyWikiWorks. This spirit is the deepest dream of all men, as it is the dream to live together in peace with all living beings on earth. And believe it or not - it is not just a dream (WhyTheWorldWorks?;-).
I think the StoneSociety in general is a good idea and can be applied practical in today's society in many places and I would appreciate such a development away from money to a more liberate system. But all these concepts will be just experimental stages on the way to complete loss of need for any regulations. (Just to say the positive things too. - I gonna contribute to the general SS discussion perhaps later too.)
-- FlorianKonnertz, ILoveWikiBecauseItIsTheForesightOfTheNewWorldOrder, 0916
Okay, it only took what, 18 years? There's a WikiStoneSociety now in private alpha on facebook. Well, it's actually a lot more than a WikiStoneSociety but this is where the thinking started. It's called DoshMosh. If you'd like an invite to the DoshMosh private alpha, email pete @ doshmosh.com. I expect to open it up to a public alpha in a couple of weeks ... exciting times! -- PeterMerel.
EditText of this page
(last edited March 24, 2012)
or FindPage with title or text search