Wcp Motivation

This is the initial argument in favor of WikiChangeProposal. Probably in bad need of refactoring.
In the spirit of CrazyThingsThatMightSaveWiki.

I came on wiki around 2000, and started contributing sometimes in 2001, but looking back in the history of WardsWiki, wiki has had its share of crises and most of them were finally settled but with less than desirable outcomes. But since the end of 2003, 2004, and now the beginning of 2005, the situation has aggravated a great deal. I do not know how others may feel about it, but from my perspective, WardsWiki is barely readable these days. Which means unusable. And that is a shame - as long as there's a chance, we could do something about it. Or if we decide we cannot do anything about it, maybe it's time to ask Ward to put wiki out of its misery, like make it read only, an outcome which many people, including me, would regret, but which would be better than the state in which wiki finds itself in January 2005.

It should be quite obvious to everyone that the prospects of this wiki evolving towards self-healing, as it has done in the past (although with ugly scars), are minimal to non-existent. Actually, that's the reason Ward himself created CrazyThingsThatMightSaveWiki. So it seems obvious that we need to SharpenTheSaw if we are to make any progress, rather than continue to fight wiki problems with the obviously inadequate tools we have now.

Ok, this boring introduction serves to justify the radicality and the exaggerated nature of my WikiChangeProposal. In the spirit of full disclosure, I shared it with Ward via private email, but Ward is a busy man so I haven't got any feedback till now, so exasperated by the miserable state Wiki finds itself in now, I decided that maybe it's better to go public, in order to get some feedback from the good friends and peer software engineers around here, and who knows, maybe we can team up a bunch of us, and even implement the software. Or in the spirit of CriticizeBluntly, somebody may convince me that this proposal is delusional/ a bad idea/ can never be accomplished/ wouldn't solve anything; please chip in. -- CostinCozianu

CurrentWikiProblems? (aka forces in the context)

Any proposal that tries to fix something has to identify what exactly it is that it's trying to fix. So assuming a minimal common general background of HowWikiWorks, and the expectations shared in common about wiki, I'll try to enumerate what it is that prevents wiki from working.

In one short sentence, the problem with this wiki, and other wikis, is that it works on a broken AttentionEconomy. Whether we like it or not, whether we agree in principle or not, the wiki has become de facto a place where fighting (ahem competing) for attention benefits is the major driving force behind wiki activity and its problems. The market is broken and if we fix it, I think we stand a chance to succeed. Some people may prefer a wiki with a more idealistic approach, but the thing is: it's no longer the case for some time, and if we face the reality of our AttentionEconomy, we may be able to reward the idealistic contributors as well.

Here are the things that break the AttentionEconomy marketplace:

Now the solution to fix the problems (1)-(4) is well-known and time honoured and tested: verifiable real identities, moderators, edited content. For example, after being the forum for XP in the beginnings of XP movement, WardsWiki lost this status in favor of forums like UseNet or Yahoo mailing list, media which have been traditionally judged as inferior to wiki. They may be inferior on a lot of grounds but they tend to address the above drawbacks quite well.

The thing is that by applying immediately identities and moderation we lose another quality that the WikiCommunity cherished even as it is in direct conflict with anything that tends to address the (1)-(4):

So my presumption is that there's a prevalent opinion that whatever solution we devise to address the (1)-(4), we should not lose the freedom. The problem is that there's no easy way to solve all 5. My CriticalSpirit also tells me that taking 5 as an absolute is not quite wise but idealistic, maybe foolishly so. But as long as there's a chance to solve all 5 points, it is worth a try.

I put them all here, to make it clear why I devised my proposed solution the way I have with all the drawbacks that it implies from programming effort, to the complexity of some operations and interactions that depart from the wonderful simplicity of WardsWiki. Although, for the majority of wiki users, wiki should still be wonderfully simple, there's an inherent complexity and effort that needs to be assumed by part of the wiki community in order to make things work. But since many good people have generously donated their time and efforts to this wiki hobby anyway, I think the rewards could be enticing.

Proposed solution

We shall have several wikis, or more precisely: several views on the same wiki.

First we maintain one wiki exactly as it is now: total editorial freedom, conflicts, EditWars, WikiNow, chaff and everything. This will take care of point (5). But, this wiki, let's call it RawMaterialWiki_ shall be transformed into an extremely unattractive prize for the purpose of AttentionEconomy, to address the problems with (1)-(4). How to do that: to begin with, we shall disable google indexing on it. And then the existence of the other (hopefully higher quality wikis) will make it less read. Its purpose will be to be dedicated to social club functions, chit chatting, hosting unpolished ideas in raw form, and constitute the material from which the other wikis can be derived.

Second. From this RawMaterialWiki_ we can set up teams of moderators (former WikiGnomes) or editors, and those will be strictly persons with real, verifiable identities. Those teams will be set up freely; any group of people, including one can set up an editing team. The function of these teams is to transform the raw material into edited versions, and even competing versions at that. Every team has access to the raw material, posters are encouraged to post to the RAW EDITION, while editors just like wiki gnomes now, do polish the material and make an edited version or a "filtered view" of wiki. Contributors can also publish specifically for a team, and other teams can only cite such contributions under "fair use". Most importantly formation of teams is absolutely democratic. Any group of people and even just one person, can make a team. This will ensure free competition of ideas. Anybody who has a gripe/irreconciliable divergence with all the existing teams (can you think of such a person? ) will be able to start a new team to promote his version of the story the way he sees fit.

Third. WikiReaders? users will be able to view any version of Wiki. Either the raw version or the preferred version. To economize the effort, competing editing team would be able to share versions of pages between them. When users type in a page URL, they will see, let's say at first the raw version, then they'll see in a nice side bar that this page has different versions (viewpoints) on it (say one view of the "lisp weenies", another view of the "smalltalk weenies"). Users will also be able to opt-in for an edited edition of wiki, and even rank the editing teams, so that when they type in a page or go to RecentChanges or QuickChanges, they'll automatically have that view selected (with a sidebar giving access to alternative views, if any).
This sounds similar to enough to how WikiPedia already works. I propose a group of you start a new book on software patterns, etc., at the Wikibooks project (or else start your own new wiki), and cull what good information there is from this wiki. Ward can then add Wikipedia or this new wiki as a sister site, or just shutdown/lock this site entirely, since it will be redundant. The only one who can fix this wiki is Ward. The only way you can fix this problem is to make a better wiki than this one. WikiPedia is already there; use it. And those who want to keep adding pages here like BS or AboutCompostorium? can stay here.

Resulting context

The crucial assumption I make (again, it may be just my wild imagination) is that such a scheme would be welcomed by the Wiki Gnomes who would like to take pride in editorship. If a team of wiki gnomes is so good, they'll be able to attract wiki users to them first and foremost. They will benefit in the AttentionEconomy.

We may even set as a prize that the "default" view for users without editorship opt-in will be not the raw material view, but the edition that was the most popular the previous week or some such form of FreeMarket competition. Under such scheme, the authors will also be rewarded because somebody say like RonJeffries or WardCunningham will know that with TeamA which is the most trusted on Wiki, his word will have more weight than a personage considered damaging's word, without him having to enter an edit war with that personage - the editors will simply eliminate the personage's "contributions". As such, valuable wiki contributors will know that their work and the quality of their effort will not be undermined by either the naive ignorance or the bad intent of any wannabe.

Under such a scheme, different points of view are welcome without forcing the community to seek a premature consensus in the document mode because the user will type something like StaticTypingVersusDynamicTyping? and what will be seen will be one version of the page (say the most popular), but immediately and prominently a link to the alternate take. Because there are two profoundly different points of views on the issue, and while we're still trying to make some progress on software engineering, it is premature to know who's right and who's wrong, so under the initial assumption of wiki, that we work towards a consensus DocumentMode, it is very hard to achieve. Well, it worked initially while Wiki community was very uniformly biased pro Smalltalk, patterns and XP, but if we are to enlarge there are equally valid points of view, and the best solution is that both should be expressed as strongly as possible, and let readers decide whom to read or whom to follow.

Also, if such a scheme is deployed, we can let somebody like Topmind, who may be a bit naive, but some of whose points may still be valid, and whose stubbornness can help his opponents objectify and refine some of their points, well, we can let a guy like Topmind wonder around as much as he likes. If he proclaims himself editor in chief of Topmind school of wiki, users will probably not bother to read him a lot, and he can repeat the same point in 100 pages and all that. Whatever valid points he has, those will be accessible to gnome editors and will be incorporated into different edited versions. He'll understand that if he wants his point to be heard, he has to raise a bar and follow some rules, so that the better editorial teams are enticed to incorporate his point of view, and his point of view be thusly heard. This is a much more reasonable approach as opposed to what happens now, where he repeats his simple-minded Foxpro point of view across hundreds of pages, thereby devaluing those pages, but is enticed and rewarded for that behavior because his point of view will thus be heard, while Wiki Gnomes are reluctant to censor him by a simple deletion where it is the case. The fact that we have a unique namespace, and a unique QuickChanges/RecentChanges, and a unique space that is indexed by google, well, all that rewards less than stellar behavior; under my scheme, that won't be the case any more.

The same goes for never-ending and useless wars on subjects like QuantumMechanics. The wars will no longer have a point when different point of views (and the egos behind them) no longer have to compete over the same unique wikispace (namespace+html page). There can be RichardKulisz's personal view of QM and somebody else's view. If RK is ranked by a reader as a better trusted editor, then when that reader types in QuantumMechanics, RK's view on the subject will appear and the other will be referenced in a side-bar. Problem solved.

As a user of this new wiki, I imagine I'll set my default options to follow 1 or 2 editorial teams, and from time to time I may look at QuickChanges for the raw material wiki. Trusting the wiki gnomes around here, I will know that my attention will not be wasted by all kinds of junk.

In the worst case, I may sign some page editorially under my own name if I feel that existing editorial teams just "don't get" my point of view and don't reflect it accurately, facing the risk that if nobody trusts my expertise in relational databases or other such stuff, it'll not be read (and I'll both waste the effort and make a fool of myself). So I predict (hope) it will work like an AttentionEconomy.

The more value an editorial team and the contributors put in the pages, the more likely it would be that their points of view will be read. At the same time, less than stellar behavior will not be rewarded. We can let RA and all others wander around making a mess of the raw material wiki, because these will not likely get them the attention they have now if at least one good edited version of wiki is available. At the same time, Wiki will still be at least as free as before and with the low barrier of entry that makes it so attractive, but you can exercise your freedom at your own risk that of not being read and wasting your time for nothing if you do not strive to provide something of value.

Please join in with comments, opinions and questions.

Programming effort

One other thing: this proposal may require quite a bit of programming effort. But the effort should not be a problem. Every year, the ICFP conference holds the IcfpProgrammingContest with problems of an amazing difficulty that are solved in either 3 days or even 1 day (the lightening division). If Ward decides that some extensive software rewriting is worth pursuing, I would like to issue a challenge: WardsWikiProgrammingContest?. Hackers from all over the world (including me) will compete for the challenge. The challenge may also include a perfect migration from the specification of the currently existing data. So programming power, I don't think should be a problem, or if there is, maybe you should discontinue Wiki altogether as will be proven a forum where less than competent programmers try to brag and give advice to others.


Some of the problems that we are facing now have also been discussed in relation to WikiPedia. See, for example, http://www.kuro5hin.org/story/2004/12/30/142458/25, for an analysis of similar forces at work in relation to providing ReallyValuableContent on WikiPedia. The conclusion that an edited version that should overcome the anti-elitism, or lack of respect for expertise seems inexorable. Wikipedia responded to this criticism (in my opinion unconvincingly) - http://en.wikipedia.org/wiki/Wikipedia:Criticisms.

Later I found somewhat similar proposals for the Wikimedia software behind the WikiPedia. See for example: Also, the comments of a brief visitor on wiki, GeorgeBrower:

There's good stuff here, but it is like many potentially good but uncontrolled on-line resources - there are way too many externalities (that's economist speak for behaviors whose cost is born by people not directly involved in a transaction - like air and water pollution). There's enough of that stuff that's unavoidable to stick around it voluntarily - a little like wandering barefoot in a chemical dump.

See also TragedyOfTheCommonsHappenedHere. The current proposal tries to privatize the AttentionEconomy so that we can fix TragedyOfTheCommons.

However, this privatization will not make Wiki less public than it is now, as Wiki will still be absolutely open for both reading and open for contributing, but there will be a privatization of responsibility which will allow for a real competition of viewpoints, if valuable versus junk, avoiding the confusion of values that is currently the result of throwing everything and the kitchen sink in the same bowl of soup. I would call this privatization privatizing responsibility. Without privatizing responsibility, I conjecture that no open forum, including WardsWiki, including WikiPedia, will ever be able to achieve or even attempt any level of excellence.

Open question and suggestion:

How does the community (hereby defined as those involved anywhere in the entire proposed wiki system) ensure that an 'edition' stays current?

Personally, I'd like to see the edited versions as branches in a version control (kinda-sorta); indeed, one could allow anyone to post (commit) to any branch, if the functionality is there to say "show me the head of all of CrazyCostin?'s commits on this branch." Almost like the editor (in this case, think newspaper editor rather than wiki writer) has the ability to bless any edition, and therefore become the head of his logical branch.

And not to drag other issues into this; believe it or not, I've actually seen a proposal quite similar to that suggested on this page (although broader in scope, which makes the whole thing trickier). Well, in all honesty, several. But YAGNI shall apply for now.

"The challenge may also include a perfect migration from the specification of the currently existing data." This, I predict, will be the great challenge of this effort, as it was of all similar efforts. How do you make the migration as transparent as possible, and yet still achieve something by it? How do you overcome the network effects to make the transition? I have no doubt that a perfect (or nearly so) migration is possible; I also think that it is vital to the success of the project.

-- WilliamUnderwood

Well, the migration can be simple: in the beginning all existing data becomes the RawMaterialWiki_. The question is whether we change the RCS, whether to preserve the versions or not. I do not understand the difficulties you're referring to.

I suggest we collect all the ideas for adjusting wiki, list the pros and cons with sections for discussion and then take a poll. There would be a place for named voting and anonymous voting. Somewhere around here is already a list, I just forgot the topic name.

This page was lengthy initially, and has grown very long. I wouldn't support anything which isn't simple enough to be explained in just a few paragraphs.
In addition to the insurmountable and actual practical problems with the "principles" (of the present wiki, where there exists problems in theory, It is also true that): One's signal is another's noise. One's politeness is another's bullshitting, one's CriticalSpirit is another's impoliteness. WikiChangeProposal provides a logical and structured approach to accommodate a wide diversity of conflicting points of view. And that was sorely missing when past participants had the best of intentions but they left disappointed. So from this point of view, Wiki has already failed to work in the past under the best of circumstances, much less can it work currently under very adverse conditions.

Q: Speaking of actual and practical problems, let me put on my questioning hat: Because the WikiChangeProposal is a good idea, and it will not "fail to work ... under the best of circumstance ... as well as under very adverse conditions" ... certainly you can point me to, show me where, it has been, or will be implemented? What comparisons and statistics will be the means of proving it is a success? If such a site does not exist, do you know of any plans to establish such a site, and who can be involved in its establishment? Just how is it to be set up, and what kind of controlled circumstances and conditions will be placed on those who would become involved? How would you populate it with pages. Where would they come from initially? Who would bear the cost of the host machine and its hardware and software, and how would its operating costs be covered? Would users be required to register and login for each session, or will the access be open and allow anonymous entities?

A: Now it wasn't implemented anywhere. The reason to presume that is worthy is that it addresses the problem of the current crop of wikis (as identified both on WardsWiki and in other communities, as per the references provided), while it cannot be worse than the current software because it subsumes its functionality. There will probably be no objective comparison and statistics, just the subjective satisfaction of readers with the quality of content created and of contributors with the quality of interaction and collaboration. The wiki would operate no differently than the current wiki, presuming that a team of programmers would implement it as open source, Ward might consider migrating the current one to the new version. Users without login session will not be able to act as editors but will be able to contribute freely to the raw material version or to any particular edited versions. Users without login session would also not contribute to establishing the popularity rank of the edited pages that they visit.

I'll think of some more in a little bit, BearWithMe?.

Basically I see two problems: you have developed a complicated plan where a lot of people are doing a lot of work according to your plan - I never saw this happen, it's against experience. The second problem is: you've no consensus on that plan and probably you will never get it because you didn't care to separate it from your ego. So people won't accept the plan if only for the reason that it's yours. -- HelmutLeitner

Thanks for your input, Helmut. Bear with me while I try to clarify a few points. <begin of long reply in version prior to my 02May edit>

Building the consensus to support this proposal is needed but postponed until I'll know better. This is at "request for comments stage. Depending on my personal and family time, I'll hopefully have a software implementation soon, but I wanted to get more feedback on building the right software. And of course the above is implied open and free (public domain, etc) so if anybody can get to implementing it faster, I'll be grateful. My ego maybe here, but my ego tells me I should not bother the community to adhere to something is not entirely clear for myself. Surely enough, I kind of signed the proposal, but my signature is not for taking credits but for taking responsibility, anybody who helps me with feedback/implementation whatever will take credit at least as much. One possibility would be to partner with someone and create a wiki dedicated to religion/philosophy/politics/economy (the most challenging topics to manage well), and see how it goes, I'll make sure RichardKulisz will be invited. But in any case it seemed premature to me to actively seek consensus for something as important as WardsWiki, at this incipient stage.

Now with regards to the amount of work. There will be a considerable amount of work in developing the software to support this scheme. But with regard to the actual work required by wiki actors it's not in my plan to torture lots of people with lots of work, as I know such a scheme would never work. That's why by default (which means for the bulk of wiki activity s) the new wiki would function just like the current one, with no more and no less work.

The more complicated editorial mechanism will be used only as necessary for dealing with issues of contention. Will this kind of work get done? I tend to think that it will, because it gets done currently in the form of dealing with trolls , edit wars, inexperienced users, etc, but with poorer tools and with worse results. If WikiGnomes and other actors had enough energy to take wiki through all kinds of wars, pests and other internet related phenomena for all these years, I have reasons to think that the editorial tools will simplify their life, and make them more effective.I'll try to suggest some UseCases for dealing with contention under WikiChangeProposal.

But under assumption that the editorial mechanism is effective (albeit effort intensive) in dealing with contention, we achieve a lot of improvement:
Costin, I have a few questions:

Sorry my experiences limit the quality of my questions, but a lot of us are like that. Thanks for considering these questions anyway. -- DavidLiu

Observation: This system curtails the RightToLeave. If I decide to retract some or all of what I've written, or even a signature, I can do so in the 'raw' content but I may have little ability to control what is seen in other views.

Let us take a look at some of the details as discussed on the private, read-only, moderated membership group called the wiki channel: I don't get it: This post is talking about how the wiki does not work and isn't readable but I've just discovered it a week ago and cannot stop reading. There's so much valuable information. What's wrong, again? I must be missing something. But, perhaps it went in a direction you don't like? If someone asked me what WardsWiki is, I would answer, "a wiki that mainly focuses on the philosophy of wiki."

The current Wiki acquired material from 2000. Most of the interesting stuff happened before late 2003, when it started to go downhill fast. So it is possible (even likely ) that readers will still find valuable pieces of information on this wiki. However, the proof that it doesn't work is that there's no new valuable material about programming added in the last half year. This assertion can be easily disproven by pointing to a page that teaches something useful and insigthful to the average software engineer and where something of substance was added after December 2004.

I agree with you there, I guess...I thought the wiki was about the wiki, not about programming. But, that's the nature of the wiki and the users decided to take it that way. That's the information that I've been finding very useful and why I thought the wiki worked. I can understand, though, how disappointing it would be to see it taken in a direction you don't want to go.

No, you don't quite get it. This wiki is about programming and programmers, it says there even on the front page http://c2.com/cgi/wiki. The "wiki on wiki" staple of the wiki world is MeatBall, if you want to learn about all aspects (implementation, social aspects, culture, interaction patterns, community building, etc) you go there. Now in the last 6 months even WikiOnWiki was affected, and I doubt anybody can point to some WikiOnWiki of any value that was added.

I agree, that WardsWiki has been going downhill somewhat sometime, but I think that was a temporary episode of WikiHistory (partly due to spam). On the other hand, I have seen quite a lot refactoring and improving recently. OK, few new stuff, but an overall improvement instead. -- .gz

>> No, you don't quite get it. Well, I am starting to get wiki. I've never visited the front page. Why should I when all the content for me is available inside it? This wiki to me is something different than what it is, or was, to you. You can say it's going down-hill, but I think it's going in its own direction that the users want it to go. That's the nature of wiki. Until more people like you come and post on programming and creating and editing worthy pages, this wiki in my experience is about the philosophy of wikis. And the phenomenon that you have disdain for is just part of the natural evolution that, although you don't like it, I don't think it can be stopped without activism. But, I have no use for the programming section and apparently a lot of other people as well, so you can see the problem. I understand. I read your WikiChangeProposal. Can you give specific examples of where it's going down-hill? The wiki, I guess, is whatever anyone gets out of it. To you, it's the programming pages, to me it's the discussion of wiki philosophy. The content has been very valuable to me.

Well, anybody can claim anything he or she wants, including that C2 is beneficially going into another direction, but you still don't get it. To tell you quite bluntly, the natural direction where it is evolving is nowhere fast. In the end, a value-less wiki will dissolve no matter whether its users want it to be about politics, bird-watching, or "philosophy of wiki". Ward himself, who is less than happy with where wiki is now and said so publicly, may stop paying the bills if all programmers leave and all that remains is a platform for wannabes to get some kind of public exposure for their useless ramblings. Obviously you have an interest in misrepresenting the reality of this wiki, but it ain't gonna change much.

If you don't want to be disappointed you can find yourself another wiki. This one is going to be a good wiki about programming or is not going to be at all. This is not a wish, it is a prediction. WikiChangeProposal is just in case this wiki doesn't make it.

I'm interested in Ward's comments about it, can you post a link to them?
I think this is a great concept. I don't know what implementation approaches Costin is contemplating, but it would seem that the technologies of the SemanticWeb - Rdf/XML or Rdf/N3, OWL, and similar material - are very relevant. This approach capitalizes on a capability of ANY hypertext, a capability that has so far been largely latent within the wiki community - the ability to construct multiple webs on top of, but separate from, an existing corpus.

Multiple views that preserve semantic integrity are *hard* to maintain. One example, the dreaded source code versioning branching and merging timesink. -- cristipp

One implication that this approach requires is the existence of some sort of write-once store. As the multiple webs are created that reference the RawMaterialWiki_, they need to know that the pages they reference do not change. I think this implies a HistoryPages-style versioning mechanism, so that one of these links is anchored to a particular version of a RawMaterialWiki_ page. We have learned, with EnvyDeveloper, that this does not need to restrict changes to RawMaterialWiki_ - if anything, such a versioning mechanism enables such changes, because they have less impact.

As a frequent critic of CostinCozianu, I vigorously applaud this contribution. I hope that I will be able to contribute in some fashion as this exciting approach matures. -- TomStambaugh

Thanks for your comments, Tom. The hyperlinks can be un-anchored in which case the user will be taken to the "best-match" function that takes as input the page name, user's preferences and possibly the current state of the wiki but in the first version it's always the raw materials. Or they can be anchored to a particular editorial team and then the user will get the most recent version of the page from that team, or they can be further anchored to an editorial team and a version number. Currently it's not implemented but I was thinking of purging all versions older than say a month, except for what editorial teams marks as "milestones". Because user should be able to reference a point in the past, otherwise I've seen the ridiculous scenario where academic papers cite "c2.com/Scomething as read on October 17th 2005".

The next thing I was trying to convince people around WikiSym is to try to have structured data on all wikis, at least just have it structured to begin with, and then make it addressable and queriable from outside. Because if wikis will be truly successful, we'll end up recreating the Web, say Web2 which is exactly the firt web with hyperlinks and everything plus the edit page feature. So then people will start complaining "we put so much data on the Wiki Web2 in plain text and now it's lost." Then a committee will be formed and they'll invent XML again just as with w3. But I'm having little retraction here, the main problem that I perceive is that everybody has his little pet syntax (like [[]] for links, || for tables and all kinds of ad-hoc solutions that are highly incompatible across wikis. With a very notable exception: two folks from Portugal created wiki for software documentation that uses XML for structured data http://www.oopsla.org/2005/ShowEvent.do?id=443. -- CostinCozianu
See also PublicVsPrivate, WikiWithAuthority, WikiStoneSociety, CollectingWikiGems, ChangeTheCommunity

JuneZeroFive (repeated interest; was started earlier)


View edit of March 29, 2006 or FindPage with title or text search