Larry,

Thanks for clarifying your original posts and rebutting some responses. I didn't realize at first that you were talking about forking Wikipedia. That would be, as you say, a horse of a different color from changing the way in which Wikipedia currently runs.

First let me reiterate what I think you are saying, so if I am again missing the point, at least you will know what I am responding to. You want to have some articles taken out of the domain of the collectively editable, and put into a space (say ExpertWikipedia) where they are maintained exclusively by some expert or experts. This would break synchronization with Wikipedia, i.e. there would be no automatic transferral of information in either direction. If a Wikipedia article evolved in a way the expert disapproved of, s/he wouldn't incorporate the changes. Similarly, if the expert article didn't satisfy members of the Wikipedia community, they would modify it and evolve it into something different. In brief, the project would be forked.

Pursuing the free software analogy futher, there used to be much talk about inevitable forking, and the potential of forks to kill off the movement. In practice, however, forks have been very rare. The only serious infrastructure fork was over libc/glibc, and that remerged before long. It turns out that two versions of the same program can hardly ever survive. Either one draws all the developers and comes to dominate the field, or all the good ideas from one are merged into the other, which obviates the necessity of having two.

My question for you, Larry, is why the same immense pressure against forking wouldn't also apply in the free encyclopedia movement. If there is a Wikipedia/ExpertWikipedia split, why wouldn't whichever is dominant gain all the momentum and all the contributions?

Suppose that ExpertWikipedia becomes the site that everyone uses as a reference work. Suppose that my contributions to Wikipedia are not being incorporated into ExpertWikipedia. In that case, I will become frustrated and quit. If I contribute further the free encyclopedia movement, it will take the form of trying to influence the expert in charge of the "official" version of the article. I will submmit my patches (edits) to her/him at ExpertWikipedia instead of wasting my time at Wikipedia.

Conversely, suppose that Wikipedia becomes the site that everyone uses as a reference work. Why would experts want their work to languish in obscurity on ExpertWikipedia? They will either quit and go back to writing scholarly books and articles, or they will wade into Wikipedia and try to get there stuff to stick there.

You talk about experts in this way:

Someone, or a group of people, that the best minds of the world can look to and say, "This is fantastic. They want to do this? I want

to

be part of it."

There's the rub. How can ordinary people be a part of the expert project? I think you need to spell out in more detail how you evision a back-and-forth flow of information between the forked projects. What would make it different from the Wikipedia/Nupedia distinction that exists today? If an expert-led free encyclopedia is such a great idea, why isn't Nupedia taking off by itself?

(7) Fred Bauder was right to point out that a lot of the people who could help *Wikipedia* most just won't put up with arguing with

people

who they think should be sitting down and taking notes.

Very true. And many expert programmers do not suffer fools gladly, and do not enjoy the frequent heated discussions which open source software projects generate. Those experts work for Microsoft, found their own startup companies, or work in academic research where they can do their own thing.

On the other hand, many academic experts *want* commentary and contributions to whatever they write. Participation in conversation is as big a rush to established scholars as it is to schmoes like me. Of course nobody has much time for unsubstantiated fringe opinions and idiotic assertions, but I'm talking about perceptive questions and critiques. I can imagine many situations where a pool of informed, cooperative amateurs would not be a hassle for a leading expert, but rather a positive draw. And one thing I, an informed amateur, can do to make an expert's life pleasant at Wikipedia, is to spare her/him from the hassles of reverting vandalism, answering easy questions, etc.

It so happens, though, that as the movement has grown in stature, those people who make the decisions really *are* software experts. [...] This doesn't contradict anything I said, moreover.

Quite so, the leaders of the big free software projects really are experts. But it does contradict the spirit of what you said, because almost none of the free software projects are "experts only", and an insignificant few projects have experts-only forks. (e.g. Netscape is a fork of Mozilla, but Mozilla is carrying the flag and drawing all the volunteer participation. Netscape is technologically insignificant; it matters only as marketing.) In almost all cases literally everyone can submit incremental changes (i.e. patches) to every project. If there is a mechanism for ordinary schmoes to submit incremental changes to the ExpertWikipedia in your proposal, I missed it.

The robustness of open source projects is vastly enhanced by the fact that anyone can contribute to any extent they like. Some of the expert leaders have worked their way up through the ranks by submitting numerous small patches, then maintaining a subsystem, then taking over entirely when a leader steps down. All this happen with no recruiting and no official designation of who is expert. It is allowed to happen because there is no distinction such as you are proposing.

I can imagine Wikipedia evolving to the point that we semi-officially designate subsystem experts (e.g. Axel Boldt as math czar), and maybe give them power to protect a small number of pages. But somehow ordinary folks have to be able to get their oar in or the project will suffer. Even the main page, which we decided we had to protect, has a talk page for making suggestions which our 39 administrators respond to.

The disanalogy between software and encyclopedia article writing is simply that software has to work. It has to do what it is supposed

to

do. As software grows in sophistication, this requires huge amounts of expertise. But encyclopedia articles do not work or fail to work;

I agree that there is some disanalogy there. But how significant is it? Please note that open source projects often exclude submissions that work perfectly well. The grounds of exclusion can be that code is unmaintainable and/or difficult for peers to review. "How well does it work?" is in fact sometimes less significant than "How beautiful is the code?", as witnessed by incomplete patches with rough edges but beautiful underlying structure trumping crufty code that gets the job done with no errors. Deciding which code is worthy is far from black and white.

Anyway, the disanalogy only matters if the quality of encyclopedia articles is not generally recognizable. It will only harm Wikipedia if people routinely fail to recognize the scholarship of experts who know more and write better than they do. So far I see some of this failure to comprehend on Wikipedia, but not as much as I suspected, and not enough to put a systematic brake on the success of the project. In fact, I think it would be very instructive to visit the discussions of some open source projects, and see whether various proposed patches generate more or less disagreement than edits on Wikipedia. I suspect you would find as much heat in the open source movement as in the typical edit war on Wikipedia.

Peace, -Karl

__________________________________________________ Do You Yahoo!? Yahoo! Finance - Get real-time stock quotes http://finance.yahoo.com

Larry Sanger wrote:

"There aren't many bona fide experts, leaders in their fields,

involved in Wikipedia right now." For example, I am not a bona fide expert about much of anything or leader in any of my "fields."

What precise criteria would you assert demonstrates a bona fide expert?

(3) My contention is that, for Wikipedia to succeed, we need experts *guiding* the *free encyclopedia movement* (notice the key words).

I notice *guiding*. It has been my experience that bona fide experts are best employed in what they excel at. Are you proposing we find some expert project managers, executives, project team leaders, negotiaters, technical writers, etc.? If so, then they could apparently apply some expertise proposing *guidance* for the community's ratification.

What I *am* saying is that, in the long run, unless a lot of experts are involved and unless there is a process that holds *some* portion of the free encyclopedia movement (not Wikipedia) up to extremely high standards,

The existing standards are as high as any individual chooses to apply.

I found the standards at infanticide extremely rigorous. In fact, I have not yet found an independent credible source to support strong claims that some of us wish to make there.

the overall project won't succeed in producing a credible encyclopedia.

Credible to who? Specific criteria can be useful in discussion.

In some cases this might be because no one but an expert would be able to write (or rewrite) an article on a topic properly. In many more cases, it will be because no one but an expert will be able to edit, supervise, and otherwise whip into shape articles on subjects that many nonexperts think, but mistakenly, they can write adequately about. There are many such subjects, at least if we want to compare ourselves to actual reliable encyclopedias.

O.K. Name a few subjects that defy writings by anyone except the leading world authority on the subject.

I contend that by the time someone is acknowledged a leading world authority on anything specific; there is sufficient written material available for the subject to be understood by most who choose to study it.

Some of the leading authorities is even more certain. Multiple authorities acknowledged implies more writings and communications and more people exposed to them in order to certify more people as leading experts.

It is insufficient to assert the Einstein (or the equivalent) is the best physicist of his generation and that only he can guide an effort to write about general relativity. By the time he has convinced enough other physicists that his theories are correct there are quite a few people around who have reviewed and written about them in various forums. These "mediocre" people are presumably good enough to contribute to quality articles at Wikipedia if they can get published in Scientific American, National Geographic, etc. or reproduce empirical results confirming the theory, or recalibrate clocks in orbital spacecraft, or read any of the above and understand it.

(4) I should have known better than not to spend at least a couple more paragraphs explaining that I do not have a fetish for formal qualifications. I agree absolutely completely 100% that it is totally possible for people who lack any sort of formal qualifications to write (and edit and code) wonderful creative works of all sorts. I also agree that this is at the heart of the success of the open source movement.

Apparently I was confused. I thought you said that the secret of their success was the prevalence of highly qualified and credentialed people *guiding* their important successful projects.

But that mere possibility doesn't mean that we don't need a lot of experts *guiding* a quality control process that Wikipedia benefits from.

Likewise, the mere possibility that leading authorities can be invited to come be in charge does not mean that we need *guidance* to achieve a high quality product.

What specific benefits to Wikipedia from a *guided* quality control process at Nupedia do you project as possibilities?

The current Wikipedia community process appears to me to implement a process that could be modeled by control theory and shown to be tending to converge towards excellence ... apparently without any pre-emptive guidance by acknowledged experts. The process seems well evolved to accept participatory input as it is available at whatever level whenever it is provided.

If the above is sustained, then as long as some experts show up interested in contributing to specific topics, and the pool of contributors continues to diversify (or rotate) then the local article quality will continue to rise as will the aggregate. Indeed, one could argue that with the evolutionary approach that has evolved the process can be *unguided* as long as the community does not fall below regenerative thresholds. It must be able to collaborate with (train) newcomers, accomplish maintenance exceeding damage accrued by casual contributors and vandals, interact a bit within the long standing ranks, and do a little original fact checking or editing. There may exist size thresholds where the overall quality of Wikipedia would drop until the reliable community of contributors grew back to previous levels and where the community would collapse rather than regenerate.

Part of

the irony in my title was precisely this point: the open source movement is full of all sorts of people with relatively few formal qualifications, and no one cares. But, IN FACT, the movement in general is guided by people who are a lot more expert in coding than the average Wikipedian is about what he or she writes about (and that couldn't be otherwise, given its success). There's nothing paradoxical about this--and it doesn't make the free software movement into a cathedral rather than a bazaar. It's a bazaar *guided* by expert coders. Kind of (but not entirely) like the stock market, a more or less free market, being guided by Wall Street gurus.

The free software movement is shaped by evolutionary pressures, not *guided*. Successful Wall Street gurus do not manipulate or guide the market as a whole (this activity is extremely localized in space time, successful "experts" get out before getting caught) they stay out of its way.

The local community has managed to develop a culture that embraces some rules and procedures which tend to evolve better material from good faith participation in the community commons, the Wikipedia. Additional participation, not *leadership* or *guidance*, will emulate and enhance past success.

Are you aware that many free/open projects start as a means of studying a language or computer science concept of interest to the initiator? Much of the expertise is developed on an as needed basis. I find it hard to believe, given this fact, that on average the expertise at project initiation that is applicable to the problem at hand is higher that the expertise available at twikification for a typical contributor at Wikipedia. Contributors at Wikipedia can apply their entire applicable expertise to small twikification opportunities and the content is immediately and visibly improved. A similar level of effort in a complex software component is likely to introduce bugs or break it entirely. Several iterations may be required before actual visible improvement occurs. Combine this with the fact that Wikipedia contributors can simply pass when uncertainty occurs regarding the potential improvement, the article does not require "fixing" before it can be compiled and evaluated by the next potential contributor. Clearly Wikipedia's articles should improve faster with fewer iterations than free software components.

(5) I am not heralding the doom of Wikipedia, Daniel M., nor did I say (or mean to imply) that what Wikipedia does is futile, and I'm sorry if I wasn't clear about that. In fact, I think that, eventually, Wikipedia *will* get the loose direction (by example) it needs, by becoming an independent part of an open encyclopedia movement that includes an (also independent) expert-staffed review board. Part of the purpose of my post was to help move the movement in that direction.

We need to establish groundwork so that academia or Nupedia can take full credit for Wikipedia's success?

You have not established to my satisfaction that your proposed leaders are intended to provide guidance in how to improve our existing successful processes. Currently anyone can (and does) read or review any article. There is substantial expertise being applied to constructive criticism on talk pages and in editing improvements. This is improving both content and editorial collaborative skills.

It appears to me that you are saying we absolutely must have old style endorsement by those at the top of the academic pyramid. This makes the material magically acceptable to readers too lazy to critically assess and establish for themselves the credibility of the material presented.

Wikipedia is being assembled by a process that deletes mistakes as it finds them. Readers are encouraged to learn scholastic research, review, and critiquing methods suitable for civilized interaction with other contributing editors and readers. One of the Wikipedia products is already critical readers. Do I believe this? Is the reliability adequate for my purpose or should I check on it? Is is consistent with my prior beliefs or sources? If it is not verifiable or consistent with other sources how much do I trust it?

There are two major methods of quality management and control used in the modern industrialized world:

1. One hundred percent inspection. Defects are corrected or thrown away. This method is preferred when dealing with potentially expensive endeavers. Nuclear reactors, space shuttles, ballistic missiles, traffic lights, etc.

2. Improve the production process and use statistic sampling to drive process improvements. This is used where productivity and quantity manufactured makes it cost prohibitive to use 100% inspection. ICs, machine bolts, automobiles, etc.

3. The two are effectively combined by the economy on an ad hoc basis to suit local economic requirements. i.e. Land Rovers or Snow Cats are extensively overhauled and inspected before setting out in redundant convoys across the Sahara or the Antarctic. Individual high quality components and initial integration were probably manufactured subject to method 2 above.

Apparently rather than challenging the reader to think critically and assess the material for themselves (as Wikipedia currently does) you propose an easy immediate jump in credibility (percieved reliability?) by being able to cite leading authorities directing or guiding the assembly of the material, rather than self improving contributing readers.

You contend that only the advertised guidance of these authorities will lend the process or the product any credibility with specialists and leading experts.

You propose that we need an "expert-staffed review board". Yet it is clear that we already have this, the review board consists of the entire community of reading contributors and the experts we have already attracted to the project.

Authorizing a "the review board says" or "As a representative of the review board I say, therefore the issue is settled" is a substantial deviation of the current project procedures which are: "If you cannot work productively without extended edit wars with the other contributors then you should come to the mailing list and discuss it or stay away from these subjects or face banning."

The latter shuts down persistent problems damaging to the community. The former shuts down effective participation.

Wikipedia is clear evidence that participation can drive a successful project. Nupedia (and possibly others) apparently has yet to demonstrate that authority or *guidance* or *leadership* can deliver a quality project without the funds to pay for participation.

(6) It is possible, as a few people seem to think, that by attracting many experts *to Wikipedia* (and continuing to forget that Nupedia ever happened) will result in the sort of excellent quality I hope we'll achieve. If that were to happen, I'd be delighted. (I don't expect it to happen; see (9) below.

Perhaps we should formulate some quantitative criteria for success and begin measuring our processes. This has been done to a certain extent, but if it is perceived as important to the community for effective self management, there is a bit more that could be done fairly easily.

But it wasn't my suggestion. My suggestion was for

Wikipedians to get behind a new or newly revitalized project (such as Nupedia), officially independent of Wikipedia, that would be managed by experts.

Managed by expert managers or leading authorities in a variety of arcane specialties? I have worked with many specialists, since they have been extremely focused on specific topics many of them lack management skills typically developed via on the job experience which professional managers accumulate.

Why expect a leading metallurgist, acknowledged top in his field, to be good at assembling a presentation of metallurgy suitable for kindergartners through P'hds as a comprehensible, reliable reference source? Why expect him/her to be good at recruiting, motivating, training, and/or managing the diverse team of specialists in education, presentation, tech writing, web design, etc. needed to tackle such a demanding multi-media multi medium project?

Roll out the red carpet. Create a structure that will make the elite feel welcome to be involved in a *leadership* role. Get universities involved, and major research institutions, and even businesses--just as is the case with the open source movement.

The elite in the open/free source movement earn their recognition by coding and giving away useful software products and source code. We have created a structure here where the elite participators are recognized and welcome. Now if we can identify ways to avoid overloading the elite, while the rest of us improve a bit, and newcomer's are assimilated without undue aggravation, then we shall really have something.

Universities are involved. Unless I am mistaken some of our better credentialed contributing regulars worked for universities. Bomis is involved by at least Mr. Wales philanthropy and unless I am miskaken there are still some paid Bomis employees actively contributing, whether on their own time or paid by Bomis or both I am uncertain.

I do not know if any influence is present yet from research institutions. I used to work at Edwards AFB, which did some leading edge telemetry processing and flight test in support of experimental aircraft from time to time. Indeed, much of what I learned there regarding TQM and project management is available to the community.

(This, by the way, doesn't mean that they would set the standards for *Wikipedia*. I would strongly oppose that; Wikipedia should be self-managing as it always has been. But Wikipedia articles are open content. They might manage a different project that uses Wikipedia content, as is their right. Wikipedia would hugely benefit if this happened.)

I am not certain I would quantify the benefit as huge. Certainly the credit for the work would be good PR. It would seem about the equivalent of a fork in terms of drawing off productive contributors. Of course, if the effort drawn was concentrated with people valuing credentials over actual content and sources for the "target" audience to read, evaluate, and contribute back to; then it might be very beneficial in reducing the scorn to which newcomer's are occasionally subjected.

(7) Fred Bauder was right to point out that a lot of the people who could help *Wikipedia* most just won't put up with arguing with people who they think should be sitting down and taking notes. A college professor who has spent his life studying X would, at least in many cases, find it absurd and ludicrous that he should have to argue with someone about X who has maybe had a college course on the subject and read a few books. There are exceptions, but they are *really* exceptions, and be grateful for them. You might hate this attitude, but it's a fact of life. The free stuff movement (how's that for a name) might be fantastic and wonderful, but that doesn't mean it'll magically change this fact.

Progress! We are really delineating some of the fundamental issues involved in this revolutionary and innovative community project now.

I agree. Modern academics are completely out of touch with the old style log, student and instructor approach to self educational opportunities that produced ancient scholars and prodigies. A doctrine of indoctrination has permeated U.S. institutions to the extent that many professionally compensated scholars, educators, and specialists feel comfortable with this resort to authority. The problem, of course, is that many people find it difficult to recover a more active role in study, research, work, innovation, etc. after spending 20 or 30 years keeping quiet and being indoctrinated.

Wikipedia breaks this mold and proposes to develop contributing readers and reading contributors who verify the facts or the reasoning for themselves. Unfortunately this potentially places a large strain on busy, knowledgeable people; some of our best and brightest. We need better processes to alleviate this, not a return to gullibility or religious embrace of opinions backed by worshipped credentials. Newton gets a lot of credit for revolutionary scientific work. There are also some allegations that he single handedly slowed down science in the British empire by routinely blocking any idea that he did not agree with or for which he could not take personal credit. Good old Not Invented Here Syndrome.

I think we should encourage our best and brightest (such as Jules) to pace themselves a bit and develop better processes for keeping inaccurate, unreliable material off the current pages served to casual readers than overworking our dedicated and best collaborative scholars. 6 Billion - 37 - 200 - 2000 - 4000 leaves aproximately 6 billion people to expose to this return to the old ways. A student, a scholar, and a log ... Actually billions of students, millions (billions, assuming that no universal scholars exist) of emulation worthy examples, and an internet connection on demand between any two self chosen people, or people and forum, addressing the subject of choice.

Perhaps MIT's announcement a year or two ago to commit to open courseware freely available online should be reevaluated in the context of our conversation to see if it provides any insight. They do not propose to argue with the readers, merely to present their easily convertible course materials online for easy, free, access to any who wish to read them. They do not invite the internet accessible public to become critical thinkers and contributors, merely a reader. Last time I checked they were still largely vaporware, but some course materials are available online.

Personally, I think Wikipedia is on to something bigger. The original goal still specificially articulated is to develop a deep, broad, reliable encyclopedia. The chosen method embraced a new technology, wiki, that was in itself a revolutionary idea. What is the simplest computer technology possible to implement to allow distributed large scale collaboration via the internet? The answer turned out to be some scripts, a server, a database engine, and a simple text user interface, Wiki. The convolution of two goals threatens to provide a miniature World Wide Web which is easily modifiable peer to peer.

Adding SVG to the capabilities really has possibilities. Since engineers are indoctrinated by society to love getting paid for building things, instead of loving to design and build things. It should be possible to continue to attract sporadic participation here even if immediate forks spring up from the green space faction's desire to save representative fragments of the Amazon rain forests via distributed R&D on technologies applciable to space settlements. Most engineers will not rush to participate until success is obvious and assured.

What will 6 billion people decide to do next when they get affluent enough to ignore their leaders' *guidance* when freely developing and sharing knowledge bases via installed sunk capital infrastructure?

The free encyclopedia project--not Wikipedia, necessarily--needs these people.

Substantiating evidence? logical proof?

Needs which ones that we do not already have?

It's frankly a little silly to expect them to help us as long as

we continue to be wide open to everyone (except "24" and Helga, perhaps...) and to follow the editing policies and practices that we all know and love.

Actually I do not find partime participation a silly expectation. I expect they will pace themselves and contribute effectively as they choose, when they choose, and in ways they choose.

For example: Jules/April have twice in the last month notified the mailing list that a couple contributors were refusing substantiate their material in credible ways. This resulted in some augmented efforts by multiple people in verifying or deleting inaccurate information. Also resulted in some further modification and articulation of community policy.

It's much *less* silly to expect a number of them to join

a free encyclopedia project advisory board of some sort, made up of leaders in all fields, that would set standards and procedures for the selection of *some free articles* (not to lead Wikipedia). It's also quite possible many of them will want to get on board as active parts of the writing and vetting process--but on their own terms, not on Wikipedia. We've already seen some potential for this with Nupedia. But I think we can do better, by getting behind the notion of a project led by, well, *real* experts. Not me, but Jacques Barzun, or someone of his stature. Someone, or a group of people, that the best minds of the world can look to and say, "This is fantastic. They want to do this? I want to be part of it."

Perhaps. I remain unconvinced because I have not identified any clear arguments beyond *this is the status quo*. The people you wish to attract have it pretty good with the current status quo. Why should there be wide attraction to revolutionary concepts within their ranks?

I also fail to understand what weight you feel the consensus of the Wikipedia community will carry with these people when you feel our product, processes, and credentials are inferior to what they will allegedly require for participation.

Again, the existing participation here at Wikipedia seems to contradict this assertion. 24/200 is over ten percent. 24/2000 is over 1 percent. What percent of the general internet population have credentials documenting the expertise you desire?

Wikipedia is already sorting the "creme". Regular repeat contributors are learning to collaborate respectfully and effectively. Methods and means of reducing workload and harassment of our best contributers also seem to be evolving.

That's how academics and scientists think, hate it or not. But it *is* how they think. Hard-headed problem-solvers will devise ways to work with it, as a constraint.

Innovative problem solvers will modify the constraints.

I think this has already occurred. I have been proposing perceived incremental improvements, not contending that the existing process is broke.

(8) A few people think I misunderstand the source of open source's success.

Similarly, Karl J., I am sure the final decisions about what to officially release are made as you say they are (by whatever experts are at hand, not by the world's greatest expert about the thing). It so happens, though, that as the movement has growed in stature, those people who make the decisions really *are* software experts. If I'm wrong, please supply me with an example.

Browse sourceforge. The free software movements source of success is its evolutionary approach. Release early and often. Reuse, innovate, whatever. If you cannot get along, fork. The best fork wins. There are a lot of losers at sourceforge. So what? It is the stuff that gets kept and used that matters.

You are attempting to argue that we must start with experts.

I am arguing that if Wikipedia is alive in ten or twenty years it will have helped develop many of its own experts.

How could the leaders of kernel releases, GNOME, etc.,

fail to be experts in what they do? The success of their projects is sufficient evidence. This doesn't contradict anything I said, moreover.

This circular argument provides no support for your assertion that we simply must start with experts to attract expertise.

Many or most of the experts you cite developed much of their expertise within their projects as they successfully evolved them.

The disanalogy between software and encyclopedia article writing is simply that software has to work. It has to do what it is supposed to do. As software grows in sophistication, this requires huge amounts of expertise. But encyclopedia articles do not work or fail to work; still, very many of them *do* require the attention, at *some* point, of an expert, in order for anyone to be able to trust them reasonably.

The free software creedo is release early and often. This means that it arguably does not work! The users help fix it.

Demonstration is easy. Go to sourceforge and search any category of interest. Download it and attempt to get it to work. Much of it will not even compile. Much of it is extremely buggy. There is a pretty clear correlation that the better it works at doing something useful the more developers and users the project tends to have. This begins a positive feedback loop until its quality delivered is good enough to satisfy most of its users, at which point participation stabilizes or falls off.

Encyclopedia articles clearly do fail or not fail. Each failure is an invitation to an existing contributor or a new reader to make a minor adjustment. The human brain is the best computer design on our planet (by our standards) and each of our readers in the near term has one applied to assessment of our material.

Perhaps we should test an advertising campaign on the front page similar to the "Edit any page, boldly" focus earlier in the project. "Please make a comment on a talk page, your feedback is appreciated and critical for continued content improvement."

(9) Now to address a point that at least three people made. If Wikipedia develops by itself, without any association with any sort of expert- controlled approval mechanism, to the point where it is used regularly by librarians and referred to as a good research source by college professors, I would take that as prima facie proof that a *lot* of experts are involved in Wikipedia.

A tautology. If experts are reading or referring then clearly they are involved with the project, even if they are not the intended target market.

But this is precisely what I predict will not

happen.

However, the only evidence that I have detected so far is based upon Academia's irrational dependence on certification and credentials rather than evaluation (and improvement or constructive comment if possible) of the actual content.

The argument seems to be:

P1 The status quo is that credentials are respected, confer respect and attract respected colleagues.

P2 Wikipedia can not be expected to change this by itself.

P3 Many experts (but not all) have credentials and are respected.

Conclusion: The Wikipedia community should take immediate and dramatic action to recruit credentialed and respected experts to *guide* *the free encyclopedia movement*.

Personally I do not see how this follows. Particularly given the fact that we have some local experts in some subjects and participation seems to be on the rise.

<snip>

Axel and Lee both opined that Wikipedia might be able attract experts to lead it (hopefully not in an official capacity but due to proper respect to their expertise in their areas of expertise)

Hopefully their expertise will allow them to demonstrate the correctness of their position to open minded skeptics. This should indeed build local respect for their abilities.

This would seem to imply first participation, then emergent leadership.

Your argument seems to be that without a priori respect no expertise can be attracted. The existing community would seem to be adequate proof that this is not the case locally.

all on its own, due to the

(eventual) strength of its material. Bootstrapping, as many people have observed.

I'm willing to admit that I could be wrong; I don't have a crystal ball. But, looking at Wikipedia's contents now and comparing it to what I recall from times past, I do have to say that I'm worried. I don't think that in terms of quality, overall, it's getting that much better.

Any quantitative or objective evidence?

But I also

admit the project is still very young and no trends can be reliably predicted.

I disagree. Sufficient data is available or could be collected to reliably project some trends. What reliability do you require? Within ten percent over the next 6 months? Confident of the sign of linear or exponential growth terms? Identification of gain or loss factors?

Notice that I do not claim I could collect the data or create the projections in isolation. Merely that the community could reliably predict some trends if it wished.

That doesn't stop me from being worried, and I think you

should be too. There's nothing utterly magical about the Wikipedia formula that *necessitates* that articles *on the whole* will not reach a level of mediocrity they never excel *on the whole*.

Control engineering is a fairly developed reliable field. Fundamental approaches have been shown to work well both analytically and empirically with predictable results and reliable design methods.

Human systems are much more chaotic than industrial processes. With that caveat noted I would say that we seem to have effective community processes, procedures, and culture that tend to establish two critical feedback processes in a massively random parallel fashion.

1. Friendly dialogue between contributors tends to improve the future contributions.

2. Random, purposeful, and chaotic editing seems predominantly to incrementally improve the local edited content.

Further, with rare exceptions, when and where diminishing returns set in, people seem to branch out or find additional interests or other people to collaborate with to the limit of their willingness to participate.

Your statement above seems to rely on the assumption that new random contributors do not learn to apply editorial judgement and that therefore future improvements beyond the *mediocre* *must* come from recruiting from a different pool of contributors with a higher credential level.

Currently new contributors are self selected, presumably because they see some potential in the project. Surely these self selected enthusiasts are more likely to apply the effort to learn new methods required by the new medium, than stodgy established authorities unwilling to interact on a peer to peer basis with the other community members?

Moreover, there's a reason to think far too many experts won't ever give Wikipedia the attention it needs: it's just not a "form of life" that they're interested in and used to. It's important that we properly come to grips with this fact.

Evolution does not have to be harsh. We do not need to assassinate or convert inattentive experts, merely ignore them. As the Wikipedia improves it is likely to attract sufficient attention to keep improving. As it improves it will eventually attract the notice of even well isolated academic giants.

It is important that we do not modify a working process, project, or community into something which does not work. That way lies "The Fact Factory" or oblivion.

My experience with Nupedia makes me strongly suspect that the ablest possible contributors to the open encyclopedia project need their own project with their own rules, and that it's unwise to expect most academics and professionals anyway (I dealt with many dozens on Nupedia) to be interested in joining a wiki and contributing in that fashion.

So the few we expect to attract should be placed in authority or leadership positions?

Axel, Jules, April, Daniel, and others locally seem to be successfully influencing the community due to their track record of effective participation. Interestingly enough even "24" seems to have exercised some influence, despite anonymity and an abrasive tendency when on the defensive.

If Wikipedia gets behind the notion, it'll happen!

What exactly are you proposing that the Wikipedia community or mailing list do to "get behind the notion"?

Let me take another stab at expressing my current mis-understanding.

Nupedia has a paid professional staff that has atttempted to setup procedures and policy, develop software, and recruit leading academics to the vision of creating the best encyclopedia ever. Presumably ideas from the participating academics/experts regarding how to organize the project, create the material, review it, modify it, publish, etc. have been discussed (ad nauseum?) and some overall consensus or decision implemented, at least partially.

Meanwhile, sometime during this process an idea occurred (between you and an acquaintenance) and Mr. Wales {perhaps advised by other professional staff} agreed to fund the experiment with an open wiki, Wikipedia ... the afterthought.

Now the randomly attracted community of volunteers (seeded by some experts and paid professional effort from Bomis) has managed to self organize, is successfully creating content, and appears to be growing and/or improving slightly. Multiple views and success criteria exist but everybody still contributing seems to agree that progress of some sort is occurring.

You contend that to achieve maximum success, Wikipedia needs *guidance* or *leadership* from the same pool of potential participants from which Nupedia's initial recruiting efforts were aimed at. Otherwise we (the entire free encyclopedia effort) are doomed to mediocrity.

I think this is a premature conclusion.

More specifically, there has been little formal analysis, specification, or agreement regarding actual Wikipedia processes to date. The ensemble is clearly working but it is made of many different approaches and peoples preferences. It is not precisely clear what percentage of the community is using which policies or customs. It is not precisely clear even what rough percentage of the newcomer's, passive readers, or infrequent tweakers even bother to read the orientation material prior to browsing the site. It is clear that google is delivering hits and that some newcomer's will occasionally become contributors. The community and product utility for initial audiences appears to be growing.

To state conclusively that the existing process is infeasible (or has specific limits to specific quality or success criteria which are attainable) would seem to require a decent understanding of the existing system characteristics.

Mike Irwin