Much has been made in recent years of cooperative ventures, such as the free software movement, Wikipedia and a host of other projects founded upon the idea of a group of unrelated individuals collaborating on a project owned by no one in particular¹. It does seem to be a philosophical approach perfectly suited to our modern age, one which denies the significance of individual achievement, while simultaneously allowing everyone who wishes to crow about their own contribution. Or, to be a bit more blunt, a system which denies the significance of individual work or achievement, as well as the importance of coordination and oversight by an expert, but also tells everyone that they are an expert. It may sound contradictory, but as I will show a bit later, it is completely consistent with the rather negative view of the individual².

There are exceptions to these views among the collaborative projects, several well know collaborations are run on more traditional lines of volunteer work. For example, the many BSD Unix versions, most of which have a central body which reviews contributions and then approve those which they find appropriate. And there are some, including some I have criticized on other grounds (mostly copyright³), which operate on something of a hybrid system. Linux is probably the best known of such systems, with the kernel rather tightly controlled, while the remainder of the code is developed on a rather anarchic model⁴. But these projects tend to be looked down upon by the true believers in collaborative development⁵.

Of somewhat greater interest are those projects, most notably Wikipedia, which claim to operate on a system of equality, and yet have established various positions with greater and lesser authority ("Wikipedia?", "Now I know Why"). What makes them particularly interesting is, while they have implicitly adopted the model they claim to reject, they do so in a most idiosyncratic may, implementing oversight while embracing an extreme leveling philosophy, with disastrous results we shall examine later.

Except for the final category, we will have no interest in examining these more traditional volunteer organizations. While they may bear some superficial resemblance to the modern collaborative projects, they are much closer to traditional hierarchical organizations, and thus do not interest us. What we want to examine here are the groups which accept the modern belief that a large enough group can collectively generate something of worth without either oversight or direction, at least outside of that they provide for themselves⁶. Though, of course, they rarely offer that description themselves.

To start, perhaps it would be useful to look at precisely that issue, the way in which the collaborationists present themselves. Their fundamental ideas are of interest for several reasons, but mostly for three distinct reasons. First, because they give us an easy way to define who is and is not a member of the groups we describe as "collaborationists". Second, because examining these fundamental concepts provides the best means of establishing the strengths and weakness of their theories. And finally because the way they present these ideas, starting from true, or arguably true, statements, then distorting them, either by twisting the meanings, or through slightly inappropriate application, they go from true statements to false practice.

But before we examine those three purposes, we need to look at how the collaborationists describe themselves, the theories on which they rely, and the way in which they apply those theories to specific objectives. Until we have those fundamentals, we cannot proceed.

The basic premise is that a large assembly, made up of predominantly non-expert, non-specialist individuals, may produce output superior to the isolated work of one or more experts working in a traditional fashion. This is predicated upon several ideas. For example, the idea that a single expert may be blind to a specific error, while when working with a large group it is more likely a given error will be identified. Or the idea that producing some output as swiftly as possible, then waiting for repeated rounds of improvement to produce ever better results will generate superior results faster than waiting to produce a single, fully formed product.

As I said earlier, these ideas are often correct, they simply are applied in ways which are inappropriate, they are used to support incorrect conclusions, or the elements which are true are either ignored, or combined with additional falsehoods, to produce faulty implementations.

For example, the idea above that one set of eyes is more likely to overlook an error than a large group is quite true. In fact, it is one of my arguments for both federalism  ("Redundancy as a Protective Measure", "The Benefits of Federalism") and the free market ("Planning For Imperfection", "In The Most Favorable Light"). However, applying it to free competition in the realm of ideas is absurd. In the realm of research, the corrective for such errors is not to place an individual work in the hands of society at large, but for society at large to review and critique the finished work, which then allows the rest of society to judge the value of each position, shaping future thought and bringing us closer tot he truth. What the collaborationist position seems to forget is that in any area of research there is not one single book, one single experiment, there are many competing projects, and that competition allows improvement, there is no need to add a mechanism for competition within each of those projects.

But, some will ask, what is the harm? What damage will be done by adding yet another check? By testing the ideas before they make it to the marketplace? Which is easily answered by pointing out that such "checks" tend to make it unlikely anything of worth will make it to the marketplace. And, if ti does, the output will be identical to all the other homogenized output of universal collaboration, and this will allow for no competition in the market. Competition requires strong positions, right or wrong, while massive collaboration tends to produce bland, committee crafted ideas⁷.

That likely requires a bit of an example to be fully understood, so bear with me or a few moments.

Let us start by looking at the dynamics of consensus. It is an area with which we are largely unfamiliar, thanks to our political system, being more of a phenomenon of parliamentary systems rather than two party democracies. However, even in the US we have some experience, though mostly form non-political contexts, such as mediation, committee decisions and other situations where we need everyone to "sign off" on a decisions⁸.

In a consensus situation, unlike a majority rule, it is almost impossible for us to achieve anything other than an absolutely centrist position, likely to disappoint everyone equally. In fact, disappointment is built into a consensus, as any single individual can hold up the decision indefinitely, where, in a majority rule situation, one side can always decide to end negotiations and "let the numbers decide". As a result, in a consensus situation there are two factors at work, fear and envy. Every individual enters into the negotiations initially trying to get their wishes fulfilled, but soon they realize doing so will upset one or more of the others, which will result in no decision being reached. And so, rather than trying to achieve their goals, they instead adopt a more conservative position, trying to achieve the most innocuous of their goals while preventing anyone from doing something they find objectionable. But even that is likely too ambitious, as even the most innocuous goals are most probably objectionable to someone, and so, in the end, they adopt a fully defensive position. Or for the most part. In addition to keeping themselves form harm, they also will likely adopt an envious position, asking why anyone else should achieve any goals when they cannot. And so, even those few innocuous goals someone might have been able to achieve are taken off the table thanks to the envy of those who are achieving none of theirs. Which is the reason consensus decisions tend

The reason I mention this at all is that true collaborative projects, in the modern sense, are perfect models of collaboration. As any individual is free to make a change at any time, the product can only achieve a final form when it is objectionable to no one at all, or, in other words, has become so bland that no one can object, just like a consensus decision.

Of course, in practice, this never takes place. More likely any given state is merely temporary, and those with differing opinions will change the product to fit their conceptions, after which it will be changed again by yet another group, and so on.

In the past, I have described this situation using Wikipedia as an example,using changing encyclopedia entries to demonstrate the harm such situations cause. But this time, I think I will use a real situation from volunteer software development to illustrate instead.

Many years ago, I did some volunteer work for the FreeBSD operating system. And one problem which I found particularly vexing was their hard drive drivers. You see, when this happened hard drives were still using a combination of CHS and LBA addressing⁹, and the standards were still in flux. Or, to be accurate, they had been established in the ATA-2 standard, but they were still not being implemented consistently by manufacturers. Allow me a brief description of how disks announced their size back then (and now). The disk drive, when queried, would return several bytes of header information. Some of these values had been established from the beginning (such as CHS data), others were added later using reserved bit and bytes which had not been used in previous standards. One such reserved bit was the LBA flag, telling operating systems the disk drive used LBA addressing. However, there were LBA systems prior to the ATA-2 standard and the creation of the LBA flag. And so, to tell machines to use LBA rather than CHS, these older drives set CHS values to the maximum possible value. Seeing this, pre-ATA drivers would know to ignore the CHS values and instead read the LBA size from certain reserved fields.

And that is where my problems arose. After the ATA-2 standard, most manufacturers continued to set the CHS values to maximum, as well as set the new LBA bit, and the FreeBSD drivers, rather than using the LBA bit, just continues to check for those specific CHS values. But, some manufacturers, mostly of mid-range and cheaper drives, no longer set those values to maximum (I don't know why) and instead just set the LBA bit, and, as a consequence, the FreeBSD operating system saw my 10 GB drive (huge at the time) as 4 GB. There was a work-around, but it was annoying to have to do it every time I upgraded FreeBSD, which I did a lot as a volunteer developer, and so I wrote a patch for the disk drive driver. In my code, rather than just check for those maximum values, it would check whether the values were set OR the LBA bit was set. It seemed a reasonable solution to me, respecting older solutions, while recognizing the new standard. And so, I created the patch file for the change, wrote up a probably far too lengthy explanation and justification, and sent them to the reviewers, sure it would easily pass any scrutiny.

And I was immediately rejected.

To my surprise, the reviewers admitted all my claims, were aware of the problem, and said my solution would work,a t least would in the sense of doing what I claimed, but they also refused to incorporate it in the code. Their reasoning, which they laid out quite clearly, was that some older drives, prior to the ATA-2 standard, did not set all reserved fields to 0, and so, in some very old drives, it was possible to find LBA bits set to 1, even though LBA had not even been invented when the drive was made. By checking LBA bit exclusively, I would break those older drives. I argued with them that the drives in question were likely on their last legs, as the LBA bit had been in informal use by some manufacturers for years before the ATA-2 standard, so any drive old enough to have a randomly set LBA bit was probably both nearly worn out and of an absurdly small capacity. But, argue as I might, they stood firm, and, for the next several years at least, the LBA bit was not checked in the driver. (I haven't worked on FreeBSD for a long time, as I began tinkering with a Forth based operating system of my own devising, so I do not know how it behaves at present.)

Were they right in rejecting my patch? I did not think so at the time, and even now I think they erred. All other operating systems, from Windows to Mac to, well, everything else, was checking the LBA bit in isolation, without using the CHS values, and so those old drives would not have been recognized by any other OS, and I have to say I agree with the other OSes on that one. At some point it is necessary to adopt the standards, even if ti break some very old hardware or software. But, whether their decision itself was right or wrong, I think they were quite right in MAKING a decision, whatever it might have been. Even a bad decision was better than no decision at all¹⁰.

And that is the problem with collaborative systems. In my case, I found a problem, I proposed a solution, it was rejected, and so it was not incorporated. It may have troubles some, but everyone still knew what the OS was and how it worked. In a truly open system, I would not have asked, I would have just made the change, introduced the change, and that would have been that. Until one of those old drives broke, and someone went in and reverted to the older code. And then, upset at the change, I went in and changed it back again. And so on and so on. Which sounds fine, except when it comes time for you try to fix your great aunt's computer. She says she has the BooFar OS, and is having disk problems. Suddenly, you have to try to figure out exactly what disk drivers your great aunt's BooFar uses. Is it mine? Or the other one? Suddenly, a product with just one name has multiple different versions.

And that is the software equivalent of the Wikipedia "reversion war" problems I described in "The Failure of Wikipedia" and "Final Comment on Wikipedia (For Now, Anyway)".

Of course, that is, as I said, only one of two possible outcomes. Yes, we very often see these sort of fluctuating situations, where changes are made and undone over and over, or maybe multiple opinions result in situations where changes are made in several different directions¹¹. But, in other cases, there may be a strong personality, capable of forcing or shaming the participants into making a final decisions¹²,  or perhaps simple frustration at the unending "reversion wars" bring about a voluntary agreement to do the same. In those cases, as I described above, we end up with the decision which produces the least possible satisfaction while simultaneously avoiding each participants specific worries. In other words, a decision simultaneously bland and unsatisfying. Only in very rare instances, sometimes when a specific subject is of very limited interest to any but a handful of individuals, or in those rare instances when there is a strong opinion shared by enough people that dissenters refuse to show their dissent, in those instances a stronger opinion might surface, but even in those cases, it tends to be a very watered down version of that opinion. In other words, rather than mixing every color into a murky brown, then diluting it with gallons of water, we start with red, and THEN dilute that with gallons of water. Or, leaving aside imprecise analogies, we might adopt a general position, but any difference of opinion between those holding that opinion will be glossed over, the whole range of possible variations of that position will be concealed and we will adopt the most generic, uninteresting, inoffensive version of that opinion¹³.

But I have spent far too long on one mistake. I said some time ago that along with the mistaken belief that consensus allowed one to find mistakes a single creator might miss, they  also erroneously argued that putting out a prototype as early as possible, allowing for rapid, frequent revisions produced ideal results.

It is a bit off topic, but I find this humorous in one respect. Many who support this position are fanciers of Linux, and critics of Microsoft, especially Windows. But one of their most common complaints about Windows is the frequency with which it crashes. Yet, they advocate a system certain to do just that. If you release early and allow frequent patching, you not only will suffer many errors early on, but will also suffer regular crashes as patches are released and tested. In other words, the crashes and patches they dislike when related to Windows are an unavoidable consequence of the design model they support so ardently.

But that is not a true criticism, just an observation I have made in the past. What is a criticism is the assumption that these multiple patches will work better than thorough design and testing would. Again, we are seeing the same problem we saw with the assumption that a mass of non-experts would produce better output than organized professionals¹⁴. First, there is the simple problem we discussed above, two or more individuals with conflicting viewpoints or design assumptions could easily continuously reverse one another's work, producing many revisions but without making any progress. Second, because of a lack of central organization, it is possible for one revision to remove parts essential to another, creating code that is worse than the previous version. Or, it is equally possible that a lack of sufficient understanding, not only of the original code, but of the ideas behind each iterative change, could result in changes which produce unexpected side effects. And, as more code accumulates coded according to multiple different standards, organized along several different lines by people with very different concepts of the project, it could become progressively more difficult to even find, much less fix such problems¹⁵. Or, leaving behind programs to look at Wikipedia and its ilk, we could argue that releasing a bare bones, incomplete entry leaves a lot of room for both the airing of conflicts over contentious topics, or for those who wish to ramble incoherently to fill in the blanks with their own brand of "information".

In fact, the Wikipedia example explains rather well the problems behind the whole concept, both the revision claim, and the idea that many contributors will produce superior output. This contention relies upon one of three ideas being true.

The first, and least probable, is that there are experts in the world, who would not write on their own, whose opinions would never be heard, but for the possibility offered by collaborative projects. This seems absurd to me, for the simple reason that, were someone happy to write his thoughts out at length on Wikipedia, what prevents him from publishing his thoughts elsewhere?

The second idea is that a mass of untrained, or partly trained individuals will somehow produce better output that professionals writing for their peers, as well as responding to the works of other professionals. This seems like nonsense, as what takes place in Wikipedia, in volunteer projects and elsewhere, is exactly what happens in academe, with professionals publishing their ideas, facing the review and criticism of their peers, responding to those criticisms, and offering their criticism in turn. If this is already happening, what does open collaboration give us but the addition of amateurs? With the exception of the possibility of an unrecognized savant, this seems unlikely to improve results. So, unless one is a radical leveler who believes experts are inferior to the "everyman", I can't see this working any better than the academic process.

Finally, there is the variation on the "million monkeys" concept. Perhaps, by virtue of producing so much output, the constant revision of collaborative projects will produce a flash of genius the more formal process of academic review would not. But I dispute this on two grounds. First, genius is almost never accidental. Despite tales of accidental discoveries and flashes of insight, it is not accident that those having such flashes were already experts in the same field. Genius comes from preparation, and discoveries come from long hard work. Second, even if there were random flashes of genius, how could we recognize them amongst all the dross found on Wikipedia? Again, it would take an expert to spot such genius insight, and it is unlikely any expert would be sifting through all the babble on Wikipedia in order to find the one in a billion bit of insight. Nor can we rely upon the contributors to identify it. Everyone thinks his ideas are brilliant, and so every contributor would claim his ideas, or those of his favorites, are just such a flash of brilliance. So, how could the experts who would need to identify such genius, every figure out which ideas are worth their time and which are not?

But I have said all of this before, in my many criticisms of Wikipedia. ("One More Wikipedia Problem","Very Short Digression On Wikipedia", "Wikipedia Absurdity, Or How To Create Your Own Citation", "Wikipedia Syndrome", "Wikipedia Absurdities ", "The Failure of Wikipedia") Yet we still see many claims offered up in favor of such projects, many times by supposed "libertarians" who compare these projects to the free market, calling them a form of competition. But nothing could be farther from the truth, and so, in the space remaining, allow me to rebut those arguments, before closing by summarizing the problems with such projects once more.

Competition, in the sense of the market, consists of offering up a good or service, in exchange for which one receives compensation. The ones who do best at meeting the needs of others then end up with the lion's share of the resources. Of course, there are some little twists to this. Such as, in one case, satisfying a strongly felt need for a few may pay better than fulfilling a need many feel weakly, though in another case it may not. But, in a general sense, competition consists of offering up a solution for which one receives compensation. One who insists on producing something inefficiently, or something not wanted, will lose money and eventually be forced from competition, as production costs both time and money.

Competition in an academic sense is a bit different. One offers up an idea, and peers then review it and judge its validity. Ideas which withstand criticism well, or which are criticized less, tend to survive, as do those which best fit observed facts. An idea which is both resistant to criticism and useful in forming explanations tends to be cited often, and this "thrives". There is also an analogy to bankruptcy in commercial competition, as one who persists in promoting an idea which has fared poorly with critics will tend to lose his reputation, eventually reducing the scope of publications which will be willing to accept his submissions, and, at last, preventing him from publishing at all.

There is a problem when it comes to Wikipedia, for example. In this case,there is no real competition, as there is no measure of success, nor is there a cost to persisting in error. If i put an idea on Wikipedia, there is no way to determine if readers agree or not. There may be some debate in the talk pages, but that is a poor measure, as talk is touched by less than 1/10 % of readers, if that. Nor can we take revisions as a measure of my success, or popularity, as it is virtually cost free to revise an article. So, unlike true competition, I cannot tell whether my idea fits well with reality, or meets with approval, nor can I tell if it survives criticism. On the other hand, if I have an idea which proves unpopular, there is nothing to prevent me from continually promoting it, even if it is clearly rejected, as it is without cost to post to Wikipedia. And so, whether or not my theory is approved, whether it has been disproved by experts, I can still place it on the site again and again.

And that is why Wikipedia is nothing like competition. Competition needs measurements, it need rewards and costs. It must have a mechanism for promoting those who succeed and eliminating those who do not. Wikipedia has none of those. It is as easy to promote an incorrect idea as a correct one, a popular as an unpopular. There is simply no means to get responses, nor to translate those responses into true feedback, into rewards and punishments which will serve to promote good ideas and eliminate bad. Wikipedia is more a tool suited for egotists who wish to see themselves in print, as nothing can prevent an individual from putting forward his own ideas again and again.

Which finally brings me to my conclusion.

For such a long article, the summary is likely to be surprisingly brief. We just discussed one of the two greatest problems with collaboration such as this, the fact that it is not competition, and, without an owner to judge what should and should not be allowed -- who would then suffer the consequences in real competition -- collaboration allows the good and bad equal time, making it as easy to offer falsehood as truth. The other problem which was mentioned a bit earlier, is the problem of inconsistency. Once something is open to constant change, it will either settle down into boring mediocrity, or else, far more likely, will oscillate back and forth between various positions promoted by those with strong opinions. As such the product will be inconsistent and unpredictable, which are almost always inconsistent with usefulness.

And then there is one more problem, one we did not discuss, but which I alluded to when discussing expert opinion and amateur. That is that these sort of projects, especially Wikipedia like projects, contribute to the welter of nonsense we find on the internet every day. As I asked in my post "Interesting Question", I sometimes wonder if the sheer volume of data, split between accurate and inaccurate, makes it harder for us to find real useful information, if the amount of data makes the truth harder to determine now. As I said in the article cites, try looking up Gisors and see if you can find any history which has not been tainted by nonsense from Holy Blood, Holy Grail, nonsense treated as historical fact.

Why I mention that example is that it is the perfect illustration. I was looking up Gisors precisely to determine how much of the entry in Holy Blood, Holy Grail was based on misinterpreted historical fact and how much was complete fabrication. Unfortunately, every article I found was already rewritten based on the book I was investigating, making it impossible to determine what had been taken from history and used out of context.("Isn't History Enough?") And sadly, I fear that as more and more of our information comes form collaborative works, which are open to constant rewriting, or from websites which site those collaborations as if they were true ("Mystery Quotes", "Wikipedia Absurdities", "Amusing "Truths"", "A Mystery Quote, Several Dubious Quotes, More Boring Quotes, and One Very Bad Conclusion"), we will have an ever harder time finding out what is and is not true.

==========================================================

1. It is interesting how, though disavowing "ownership", those who create many of these projects have no problem basking in the fame their projects generate, sometimes even abusing that fame (Jimmy Wales comes to mind, see "Yet Another Wikirant"). It seems if these projects were truly collaborative, then claiming to be the founder would have little significance, basically putting one on par with the man who made the paper on which Dostoyevsky wrote The Idiot. But, as we have seen ("Wikipedia?"), many of these projects, while claiming total equality, maintain an Animal Farm-like system of some being more equal than others.

2. I am sure many in the free software movement and other projects will deny they have any sort of agenda, and will most definitely state they do not deny the importance of individual achievement. However, it doe snot matter that they do not consciously hold these positions, their philosophy is still the outcome of those beliefs. As I have argued elsewhere ("Inescapable Logic", "Recipe For Disaster", "The Endless Cycle of Intervention",  "The Cycle of Compassion", "Why We Lose", "Giving Away the Game", "Pyrrhic Victories", "You Lose When You Think You Win"), once a belief is accepted, it is taken to its logical conclusion, whether those doing so realize it or not. These projects may not consciously aim at denying individualism, but by declaring there to be no need for ownership or supervision, and by claiming anyone can make a change, essentially asserting the equivalence of all opinions, they are adopting this position, accepting a belief that is "in the air" of the modern society, even if they do not recognize its influence.

3. See "Copyright as Politics", "Examples From Another Field", "Object Oriented Programming, Apple Computers and Justice", "Some Libertarian Analogies", ""Best Practices"" and "More Examples From Another Field". These particular aspects of what I dubbed the "libertarian left" ("The Libertarian Left",  "The Failure of Wikipedia", "Revelation From Bottom Feeding") mirror some aspects of the real left, see "The Inherent Disappointment of Authoritarianism", "The Threat of Perfection", "Utopianism and Disaster", "The Danger Inherent in Banning "Bad Ideas"",  "The Right Way", "The Most Misleading Word", "Luxury and Necessity", ""It's Our Top Priority!"", "Absolute Values", "Some Thoughts on "Summerhill"", "Musings on Discrimination" and "Liberalism, Its Origins and Consequences".

4. Actually, Linux is difficult to describe with generalizations, as some distributions are commercial, and thus developed like any commercial product. Others are controlled by a central group, akin to those controlling the various BSDs. And others are, as I described, following the modern anarchic approach. But that is not the end of this description, as many drivers, libraries and applications are shared by multiple distributions, and subject themselves to a different system of control, so even the most rigidly controlled commercial distribution may contain many elements developed using an anarchic free for all model.

5. As the collaborative development advocates tend to be linux fanatics, and as linux is famous enough that they want to claim it as one of their own, most often they downplay the control exercised over the kernel, and ignore the existence of more rigidly controlled distributions (and the large part they played in making linux a viable operating system in the commercial world), and treat it as if all of Linux were developed by an uncontrolled mob.

6. Some will argue that a free government is nothing but a large group providing themselves with direction, but there are quite significant differences between a free (libertarian/federalist -- see "Reticent To Adopt a Title", "A Possible Designation", "The Right Identity", "Three Approaches to Social Conservatism"), government and the sort of disorderly mobs which create Wikipedia articles. As I will discuss this later in the post, I won't get into it here, but I wanted to make readers aware that I recognized this potential comment and was planning to address it.

7. At least this was always the historical case, and is still the case when there is a demand for a fixed, permanent copy. In some digital products, where content is not fixed and can be continuously altered, there is a second possible outcome, a work characterized by two or more very strong opinions constantly overwriting one another, a phenomenon seen most often in Wikipedia articles on contentious subjects, where competing viewpoints alternately seize the floor and eliminate all traces of any other opinions.

8. To be accurate, even parliaments are not likely to produce results as completely disappointing as true consensus decisions. A parliament is still governed by majority will. The only difference between parliament and two party democracy is that parliamentary decisions usually need to placate more groups, making decisions more like a consensus than a two party system would produce.

9. Early disk drives were made up of cylinders, heads and sectors (hence "CHS"). Cylinders were the number of circular zones on each disk platter, sectors measured the size of each pie shaped wedge on a given cylinder, and heads were the number of disk read heads, which told the number of platters (actually numbers of readable sides on the platters). By multiplying them, you got the number of bytes a disk drive could hold. CHS was limited in potential maximum size by various operating system and BIOS limitations, and had a number of other shortcomings. As a result, Logical Block Addressing (or LBA) was developed. Rather than addressing by Cylinder, sector and head, the LBA address was simply a number which was assigned to a single location on the disk, allowing for much more simple software, as well as allowing larger maximum disk sizes.

10. This argument is almost identical to the one I make in several posts that bad laws are better than unpredictable ones. See  "In Praise of Slow Changes", "Predictability", "Conservatism, Incremental Change and Federalism", "The Problem With Evolving Standards" and "Inflation and Uncertainty".

11. In our highly polarize age, this is a hard one to find in my favorite source of examples, Wikipedia, but there is one special case. Go to a topic where there are two highly partisan sides, such as, say, the history of the conflict in Bosnia. Or better, something of a slightly more "historical" nature, that is something old enough that amateur historians take an interest. Say the Greek war of independence. Generally, as you would expect, you will find two positions going back and forth, Greeks and Turks each making changes and removing changes made by the other, to make the page reflect their position. But, in the case of "historical" events (I also saw this in the pages on Gavrilo Princip -- see "The Problems of Group Identity") there will be ostensibly objective outsiders trying to revert changes made by both to achieve some sort of objectivity. And thus you will sometimes find three sided fights. (Or perhaps more, as various schools of history attempt to impose the brand of "objectivity" corresponding to their socio-political views -- cf "The Impossibility of Unbiased Reporting" and "The Death of Impartial Media".)

12. Wikipedia has two tools to supposedly avoid these problems. First, editors who can lock topics, but who, for the most part, simply impose their own opinions, resulting in reversion wars among editors, locking out some participants, but not making for any more settled content. And then the ability to substitute a vote for reversion wars. However, as each vote is public, as is the ongoing count, these tend to settle little, as anyone with a strong opinion can usually gather sufficient support to keep the debate going. I discussed them a little in  "Wikipedia Absurdities" and "Wikipedia?", but they are among the lesser problems I have with Wikipedia as a whole.

13. A good example would be modern "polite" discussions of government, or perhaps the sort of "survey of government types" you find in high school civics texts. Everyone agrees, for the most part, that representative government is "good", but what type of government, or what "representative" government entails, is usually left terribly vague. In other words, we embrace "representative government", but ignore specifics, forget all the many disputes over specifics, and adopt the most general, uninteresting, bland possible position.

14. I normally phrase this as "Wikipedia has taught us that a million monkeys typing on a million computers for a million years will still produce nonsense." Of course, Wikipedia is not exactly analogous to the "million monkeys", but there is an element of that in the claims that non-experts will produce better results. But this will be discussed shortly.

15. I have worked as both an administrator and programmer, and I can attest the worst code is not that by a single bad programmer, but code which was touched by several supposedly "good" programmers. As each of them applies his own clever tricks, his habitual shortcuts, or his own idiosyncracies (mine is using pointer arithmetic everywhere), the code can become all but unintelligible.

================================================================

POSTSCRIPT

For those who have not followed my blog, here are all of my comments on Wikipedia:

Stop Confusing Me With The Facts!
Mystery Quotes
Wikipedia?
The Failure of Wikipedia
Opinion Masquerading as Fact
Funny Numbers
What is Wrong with Wikipedia?
Endangered Species
Sterility of Formal Economics
Deceiving Themselves?
A Question About Language
Why People Don't Take Academics Seriously
A Request for Rush Fans
Wikipedia Absurdities
Proof Positive
Some Libertarian Analogies
Have to Love Wikipedia
Wikipedia Syndrome
Yet Another Wikirant
Final Comment on Wikipedia (For Now, Anyway)
Wikipedia Absurdity, Or How To Create Your Own Citation
Now I know Why
Rush Debate
Vindication
Predictability
Life is Strange
The Power of Myth on the Internet
In Defense of Standards
Addenda to "In Defense of Standards"
Best of the Web Imitates Me XX
Roman Legions, Hopscotch, Killer Gays, "Got AIDS Yet", WMDs and a "Damn Piece of Paper"
The Libertarian Left
One More Wikipedia Problem
Very Short Digression On Wikipedia

And, since it is the outcome of yet another collaborative effort, the following essays all contain (as comments at the end, in most cases) criticism of the spell checker which comes built into Firefox (In most cases, the comments on spell check come in notes at the end of the essay):

Why McCain Will Win
A Problem With Amateur Historians
A Cure For Infertility
Can I Hate America Now?
Why Nuremberg?
Short Response to Boortz
Another Useless Bureaucrat
For Your Amusement
Too Clintonian For His Own Good
A Question About Language
Spelling Nazi V
When Help Hurts
Bad Statistics
Expectations
She Won Me Over
Oh No, Not Again
Sorry Peggy, Wrong on This One
Where Are You Tom????
The Spelling Nazi Begs to Differ
Grammar Nazi Comment on Greco-Latin Words
An off-Topic Pet Peeve
Some Additional Thoughts on Technocrats
Res Ipsa Loquitur

While the former set provides quite a bit of discussion about the causes of such problems, the latter simply documents the problems I observed, in a few cases speculating on why they took place.

POSTSCRIPT II

I expect some to criticize me for limiting research to experts, but that is not precisely what I am suggesting. First, let me say that I am not an expert, so I do not make this proposal out of self-interest. In fact, as an amateur economist, I am actually harmed by the strict publication standards I suggest. And what are those standards? Nothing new. I simply propose that we treat collaboration with skepticism, and should be more inclined to favor traditional peer reviewed sources. Of course, such sources do tend to favor the experts, who know the material well and are familiar with the standards and conventions of the field, but there is nothing preventing non-professionals from submitting and being published, they simply must meet the standards applied to the professionals. (At least in journals which review blindly.) But that is a topic we should deal with another time. For now, I just wanted to point out, once again, the problems with modern collaborative projects.

POSTSCRIPT III

I realized when I finished that I had not gone into detail about the fact that this tendency toward collaborative work and projects open to all was an expression of a philosophy which rejected the concept of expertise, and denied human ability. I am afraid as my post went on, ti simply slipped my mind. (Not to mention that my post was delayed for almost half a day by a technical problem on blogtownhall.com, which kept me from writing anything for half of Sunday or more.) It is an interesting topic, though, and one I wish to examine in some detail. But I don't feel up to rewriting this post right now, so instead, I will come back in a few days and discuss the topic on its own. Hopefully that will prove interesting, as well an answering whatever questions my readers might have.