Larry Sanger wrote:
Dear all,
I've been away from active involvement with Wikipedia for many months now, though I occasionally still lurk on Wikipedia-L and make a casual edit on the website when I feel so moved.
My distance from the project, and some recent reading about Linux and open source software, has made something clear to me in the past few days: there is a profound disanalogy between the development of our free encyclopedias and the development of free operating systems and software.
Organized information is organized information; however, we have a profound advantage in that our "compilers" have human judgement and initiative.
In particular, the Wikipedia project has been defined in such a way that we have few official standards and no virtually requirements for quality of the rigorous sort that Linux had when it set out to rewrite Unix from scratch (and later remain compliant with stringent technical standards like the POSIX standard). Linus Torvalds' task had well-defined parameters that absolutely required a lot of genuine expertise. Our task, by contrast, is to write a very large, unbiased encyclopedia. What this task entails is far more nebulous (though I and others have worked very hard to settle on and explain what it does involve), and many reasonable people reasonably think that this doesn't strictly speaking require genuine expertise.
There are a lot of implicit assumptions above that, in my view, are incorrect.
The only rigorous standards that were applicable to Linux, other than what the developers themselves applied, was the goal of being compatible with unix interfaces and a c-compiler. Posix compatibility was a goal, not a rigid requirement. It is also likely that the kernel developers occasionally tweaked the compiler to get rid of irritating anomalies.
It is my understanding that Linus was an undergrad C.S. major when he launched his kernel effort. The originators of the GNU components and development environment may have had a higher mean of starting qualifications.
Clearly some expertise was developed in the course of the project by the contributing team members.
One might easily conclude from the "No original work" non policy here that sheer scholarship is perceived as sufficient to both start and finish the Wikipedia project. The required research seems to be slowly expanding my horizons, I suspect this is true in general. Personally, I do not think life experience should be discounted. Local assessment of world events often seems a bit different than the material broadcast. Old books are where you find them. Humanity keeps a lot of knowledge stashed in various places. Non academic personnel know a lot of astonishing detail about various stuff they have experienced.
But it does. If you think otherwise, you're living in a fantasy world. The fact that there is no organization like the IEEE staffed by world-class experts defining a standard that we must follow doesn't mean that our work doesn't require expertise to finish credibly.
What kind of expertise? All kinds? P'hds?
Do you think millwrights or truck drivers have anything to add regarding the forest products industry or various technologies?
Avionics techs or reactor operators anything applicable to the technology articles?
Housewives and parents to articles on babysitting, food preparation technology, or infanticide?
Reckon Michael Jordan knows anything about professional basketball that P'hds do not?
Concrete finishing.
Was Medicine Man an accurate portrayal of arcane knowledge ocasionally not yet seen in western academic settings or merely an urban legend or Hollywood fantasy?
When you say credibly you mean to academics? This is a target market issue ... "24" attempted to facilitate a discussion about this at meta but nobody seemed interested.
I am pretty certain I am not participating here for the benefit of academics. Most of them in the U.S. have access to pretty damn good libraries.
One of my payoffs was free interent distribution to any reader that desired access. Where I went to grade school some of the kids did not have an Encyclopedia at home as I did. It made it difficult for some of them to complete their homework. I understand that there are 3rd world nations where libraries and books can be difficult to access. My thoughts on this are that the price of portables is still coming down and eventually the world will mobilize to provide net access. IMHO It will be cheaper and more effective for U.S. citizens to provide tools, information and market access to allow people to develop their own economies and world wide markets than to pay taxes for the U.S. military to protect wealthy unscrupulous U.S. citizens interested in exploiting trapped and desperate labor sources.
Other perceptions may vary. Too bad we deferred that discussion but we can always start it up again.
I think
writing *and finishing* a credible draft of an encyclopedia requires more and a wider range of expertise than the free software movement has. If our encyclopedia project doesn't get an infusion of that expertise, the quality of the result will suffer accordingly, which is a lot.
So how many free software developers do we have? 3? 5?
We seem to have a substantial infusion of non software developers if the discussion on this mailing list is a reliable indicator. Certainly the experience briefs provided by various Wikipedians on their home pages seem to vary broadly outside the software arena.
I was under the impression that additional free software assistance would be welcome, not superfluous.
The problem is that, with several notable exceptions, highly-educated people aren't drawn to Wikipedia.
Disputable. Extremely. See Karl's? strategy for long term recruiting with his father, the specialized historian, retired. I have not run a detailed survey but I suspect that the community is actually over represented by percentage of educational credentials compared to the available populations with internet access.
If you mean multiple P'hd's doing original research ... duh! We keep telling people no original work. No discussion. No speculuation, get back to work! 38K and counting.
The GPL would likely deter publishing of papers that peer journals wish to be copyrighted for profit but ... we could establish forums or a policy that personal original papers contributors wish to publish under the GPL could be linked to their personal pages and protected. Or we could encourage them or a copy of them at an appropriate place in the stack, serene in the knowledge that eventually another expert will wander by with an edit boldly or NPOV or citation needed gleam in their eye. Not every essay a professional writes is suitable for top scale peer publication, nor is it necessarily worthless. Might be just the draft to initiate a firestorm of collaboration at a weak spot in the stacks.
I once saw a thesis on calculating the probability of life elsewhere in the universe linked to for comment from sci.space.policy and a couple of other usenet groups. If the original paper and resulting 300 message technical debate had taken place under an FDL on a Wikipedia type medium, much NPOV'd content could have been extracted for a variety of articles.
Not a debate forum. Back to productive stub editing and augmentation. 39K and counting. If I made a further comment here someone might wish to indict me for contempt of non policy. Perhaps our desired credentialed are debating on usenet? Think of all that I.Q. discourse recorded in lengthy linear fashion in google archives but not available for synthesis in the Wikipedia.
Not a debate forum.
One of the most reliable and widely used forms of learning in the wild, in the military, in the home, in schools, and various other places is emulation.
I am not saying that all of our antisocial ban candidates could or would learn from watching passionate intellectuals go at it and occasionally commit the odd fallacy of attempting discreditation of the opponent or his credentials rather than his argument only to recover composure gracefully and resume intelligable arguments for discerning readers. I make the far weaker claim that it might positively affect some of them who were previously unfamiliar with civilized discourse.
Yes! Yes! I know! Not an ideological forum or tutoring facility.
Still, if we cannot attract credentialed contributors, we might need to roll our own. From H.S. to college degree ... 4 or 5 years. From B.S. to M.S. or P'hd another 3 to 5 years. In field stature ... priceless 5 to 30 years. We could have home grown credentialed in anywhere from 3 to 50 years depending on the initial starting point of the candidate.
It's not surprising why not: I would
like to suggest that this is similar to asking veteran programmers working on Linux and its applications to work with, supervise, and put up with rank beginners and script kiddies.
It happens. A rapid learning gradient and sometimes a certain amount of hero worship is required. At the other extreme, only patches of significant contribution or quality are accepted. Some maintainers will put substantial effort into debugging and integrating a significant patch, others (wisely IMHO) eventually tire of cleaning up sloppy work and simply discard it until the submitted patch quality improves. (This is all hearsay, I am not personally yet a free software developer. Wikipedia is kind of a personal first project for me since the software looks like an interesting prototytpe potentially useful to some other open or free projects I have been considering.)
If they had had to do that, I doubt
very much that the free software movement would have come a fraction of the distance it has.
I think you provide an incorrect impression here. Some free developers will go out of their way to coach newbies while others avoid them. The project environment varies with each project team's attitudes and requirements. The attitude or time available varies widely from developer to developer. The attitude also varies with the effort needed and available for specific projects. For verification, substantial original source material is available from newbies and experienced developers at advogato in the front page articles as well as the personal diaries of account holders.
It is my impression that the free software movement has been somewhat overwhelmed by its current success and publicity. The ratio of newbies to experienced has shot up so fast that veterans must allocate their time between coaching and software development. This will be self stabilizing. Either the free software community will develop means and methods of orienting newcomers effectively, or they will get discouraged and move on. There are many sites and approaches being experimented with to help out newcomers. Sourceforge and savannah provide free use of free tools; advogato, and other sites are experimenting with community trust metrics, many experienced developers are posting tutorials, tips, procedures, etc. online for newbies benefit. Several large projects such as gnome and debian have established non-profits and sites to help address issues in integrating large projects via coordination of large teams experiencing substantial interest from neophytes. Finally, most projects have a dedicated mailing list or forum where advice and assistance is available to varying degrees. ... and of course the first thing to do is to read the code.
It is pretty clear that if the free software movement was not pretty friendly towards new developers it would not have grown or been successful. The economic benefits of free software development are huge to society but currently scarce/long term for developers. Although the expertise developed is valuable in the long term, the community grows because people interested in software development enjoy working within the various teams and with the individuals who comprise the overall community. There are other places to gain the expertise if that were the sole goal.
Please don't misunderstand. My concern with expertise and knowledgeable participants does not reflect an overvaluation of formal qualifications, or academic elitism, by the way. (If you think I have enormous respect for someone just on the basis of their academic credentials, you *really* don't know me.) If someone without a degree (I can think of a few) can write and think well and convey what they know in a way that reflects expert knowledge on the subject, that's great. May their kind be fruitful and multiply (among our ranks). There's no reason for me to suggest otherwise, just as there's no reason to ask free software developers to have degrees in computer science before they get their hands dirty working on open source software.
Interesting. Credentials are unnecessary but desirable?
Consider this. Eric Raymond might be correct that free software development is represented as a bazaar. What is perhaps less often acknowledged is that it is a bazaar full of extremely highly-qualified, knowledgeable people. In this bazaar, the bar to *productive* and *original* development is set very high. (Conveniently, it's not people that set the bar high but instead the facts of reality about how hard it is to develop software.)
lol Most free software is crap until it gets good enough to attract experienced developers who use it themselves or support it for their customers. What original development? The majority of killer free apps are clone knockoffs. What Microsoft has done to the planet, disgruntled free developers are free to do to Microsoft (and everyone else) until sufficient bribes reach U.S. Congress Critters to outlaw it. Further bribes to the Supreme Court, or strategic bench stacking, to find speech is free unless speaking code.
Certainly there are highly qualified and experienced people and they have produced high quality libraries, tools, etc. This does not imply that all work of value coming out of the free software movement is performed by these people.
The primary bar that I have observed is the software must interact correctly via open published standards. Thus it works immediately with most existing working software available from the community online and can be replaced later by a better module if desired. Occasionally a killer free app will create a popular standard. I think Gimp helped put png on the map. Ogg Vorbis has had a harder time, probably due to the wide free (as in beer) availability of mp3 tools and sound tracks.
It is also less often acknowledged that there
are necessarily elite groups--elites based on merit, but elites nonetheless--who are in charge of releasing new versions of important packages. That's as it should be.
Often the merit consists initially of volunteering. Debian seems to be training newcomers to maintain packages as fast as they can attract them. Gnome and KDE also seem extremely interested in help from newcomers, particularly with debugging.
The many eyeballs makes bugs shallow approach seems to require many eyeballs.
I will grant you that important packages end up being maintained competently. Otherwise they get forked when people relying on them get tired of impacts from incompetence. Although tutoring seems an often preferred option, after all if one forks then one must do all the work oneself. If the incompetent is willing and able to learn tutoring can be a highly leveraging experience in this situation.
Wikipedia is quite different. The bar to contribution is very low, and if there is any elite in charge, then with all due respect to everyone (and that's a lot--there are a lot of *extremely* smart and knowledgeable people here), our elite would seem rather less than impressive compared to the leading members of the intelligentsia that contribute to the likes of Britannica.
Less impressive to who? Personally I find our volunteers (and the few professionals getting paid for time expended here) fairly impressive. I personally anticipate that as the internet divide moves economically disadvantaged peoples everywhere will be very impressed with our free product even though and especially because it is provided to them free of charge. Spend a days pay on a used National Geographic or have the kids practice reading on U.N. provided solar powered satellite linked portables at the schoolhouse? Tough choice for a subsistence level farmer or budding internet entrepreneur.
We have some crude feedback loops in place that are working to improve both the content and the collaboration abilities of most of the contributors.
It would seem to me that the proof is in the pudding, not projected from the first draft of the cookbook.
Along these lines I suggest there's another disanalogy between the free software movement and our free encyclopedia movement. The free software movement is organized and led by world-class computer scientists associated with industry and academia.
Well perceptions vary but to me it appears not "organized" beyond agreeing widely that open standards are important, source code should be freely available and freely redistributable, interoperability and interconnectivity are king, and killer apps cannot be ignored. In fact, I would have no problem calling it chaos or anarchy. Everybody does as they please and occasionally brownian motion defeats entropy and a team is loosely bonded. This team in turn acts with greater attraction to randomly moving developers and especially neophytes. Entropy continues to lose ground as affinities between productively contributing developers grows. This leads to faster team formation next time and random anarchistic collaboration between teams and individuals.
Led? The "leaders" seem to write software and occasionally express a public opinion or post an essay. Often about guns, license ideologies, or other pet peeves. Many neophyte developers seem to start out by posting their own efforts. Eventually somebody joins somebody else at a project or idea. Communication and learning is established. Eventually sufficient skills are established that a patch is accepted to some crummy project installed on 3 machines and projected to be viewed in the history of the world by a grand total of 100 eyeballs ... mostly rolling and yelling Yuck! Yuck! Nevertheless, feedback loops continue to operate, and the neophytes start doing something useful to somebody: web site design, mailing list management, debugging, eventually some code. Occasionally lightning strikes and a neophyte with sufficient skills and social graces to lead a major project establishes a project that looks widely useful. This attracts highly experienced (sometimes even credentialed) developers and grows into a major app. I have seen it alleged that the Linux kernel was exactly such a lightning strike assisted by the fact that the entire rest of the GNU components and tools were complete and ready to use.
The free encyclopedia movement is
much newer, but (forgive me) it doesn't seem to be travelling in the direction of being led by world-class thinkers, scholars, and scientists, as a close analogy would seem to require.
A contrived strawman. There seem to be a number of distinguished panelists listed as participating at Nupedia. By your theory of leadership it would seem that Nupedia should be farther along than Wikipedia.
To be quite honest, it was good
to lay me off when economic necessity required; now do the right thing and ask Jacques Barzun (before he dies), or some other distinguished intellectual, to head up the project properly.
We need a "distinguished intellectual" to lead Wikipedia? How do you account for the success to date? All indicators seem to imply that the community is growing, the content is growing, browsing randomly seems to find more content and even more quality content than 6 months ago.
Anyway, why not send Jacques Barzun (whoever he is) an engraved invitation yourself. AFAIK there are no limitations on participation beyond a willingness to collaborate and occasionally shout back. Shouting is optional, many adoit and sensitive people simply prefer an inconspicious fade away. Personally I have put any recruiting drives on hold until some procedural stability and community guidelines emerge that seem viable for the longterm to me. I do not think we need lots of highly credentialed people to "lead" us, what is lacking is massive participation by thousands of diverse contributors and the means to manage the inevitable conflict that will arise between them. Still if the credentialed elite choose to participate then I, for one, welcome their assistance in this worthy endeaver.
If we really want to make the best encyclopedia in the world (the original stated goal of Nupedia, by the way), we must discuss a pressing question that I suspect very few people on this list are disposed to take very seriously: how can we arrange for our free encyclopedia movement to be led by representatives of the creme de la creme among the world's scholars and scientists?
Simple. Invite/Let them volunteer to come do some work. If they propose better procedures and develop a community consensus after establishing some credibility here then we will be lead by the "creme de la creme". If our current procedures are perceived as flawless then they can contribute until community omens are more auspicious towards change or they can recruit a quorum to out vote the previous cabal of regulars.
Until the new volunteers show up and dazzle us with their performance I propose we muddle on, improving on what we have so far.
Now, I would not dream of suggesting that *Wikipedia* change its policies of openness. Basically, I don't think Wikipedia should change. It is what it is and it has produced a huge number of *great* articles. It's amazing that it works as well as it does, and I continue to expect that it will result in a useful, interesting, huge body of work if we continue on in the same way we have been.
It works and success is anticipated. What is amazing about it? It is well documented in the TQM literature, differential equations, engineering science, economics, and history that small improvements add up. Higher order effects such as improving or adding personnel tend to accumulate more slowly initially as new personnel are assimilated and then the impact accelerates as improvements are applied to a larger base. As I recall I was highly excited and energized to begin contributing immediately upon encountering it, 2 - 3 weeks later I was highly irate because it appeared that I had been suckered into contributing to commercial endeaver interested in making the database difficult to fork. At no time was the viability of the project technology or approach in question in my mind. Only whether it would become/remain a free project. So far, it has become a free project and seems well on its way to success. I see no reason to suspect that all future discerning contributors will be less discerning than you, I and the other repeat contributers. Surely it must be getting more obvious all the time with improved software, community, and content.
That said, all of my previous predictions of huge success for the free encyclopedia movement were based on the assumption that a Nupedia, or some other quality control mechanism, would eventually mature into something to inspire confidence among the leaders of different fields, so that contributions and editing would be of the highest quality.
Highest quality or highest credentials?
Writing e=mc^2 seems a bit different from deriving it. Indeed, now that Albert has derived and published it; I would contend that I can type it as well as he could, were he still alive.
Are you retracting your predictions?
I am willing to predict that even with no changes to current procedures the Wikipedia project has and will inspire clones and derivative projects using derivative technology. Further, the content will be used widely, tailored and improved for a variety of purposes. Even further, academia and peer journals will be highly irritated and even threatened by our open rapid revision and delivery of high quality content.
At some point some of the credentialed "experts" or authorities will decide to join us to explore the technology and it effective uses in communicating with students, colleagues and others worldwide at no cost other than participation or donation. If they are polite about it and learn our local customs we probably will not ban or ridicule them.
If we need them, we can whip up some further wild eyed predictions and attempt to substantiate them adequately for [[Replies to our critics]]. I have not checked this lately perhaps it has gotten out of date.
But if no such
mechanism materializes, I would be much less apt to predict success, in terms of quality of articles, for Wikipedia. Wikipedia by itself will continue to go on to useful things, interesting things--but not great things.
Great things are available only to credentials? I think you are lagging the internet a bit here. Linus has been acknowledged an expert in operating systems after the success of Linux, not the other way around. In fact you may wish to review the "Tannenbaum" incident on usenet wherein a credentialed expert (with a financial interest in the competing minix) attempted to educate Linus out of his "obsolete" approach to writing a kernel.
Indeed, I think one of Wikipedia's greatest assets is its diversity. The material here is not carefully crafted by some professional writer working under an editorial policy dictated by the magazine owner. It is hashed out by all who choose to participate and, as a result, likely closer to a summary of the mainstream view of objective reality than has been seen since ..... ever. Welcome to the Internet and the future of communications. Point to point direct without the intermediary special interest groups or choke points.
If Harvard, Yale, or Mit people decline to participate then the brilliant prose will merely be from the rest of the participating planet. Oh yeah! I am worried about our future now!
So I don't propose we touch Wikipedia--but we have Nupedia. What I hope is that Nupedia can be changed and rearranged, somehow, to create an elite board of bona fide experts that is ultimately in charge of "releases" of free encyclopedia content.
Whatever the specific Nupedia article creation and/or vetting process might turn out to be--see the Nupedia-L archives for discussion ad nauseum--one thing is increasingly clear to me. Namely, unless there is a dramatic change in how the free encyclopedia movement is organized, Wikipedia will be stuck with, on balance, mediocrity.
You beg the question here. I have 15 years experience recruiting, organizing, funding, and managing project teams. It is amazing how fast mediocrity is left in the dust with even a tiny well functioning team. In fact, it has been my experience that prima donna genius is best allowed to sit on the sideline if it cannot work effectively with others within the team's procedures and constraints. It does not mean genius is ignored, merely that integration proceeds with a minimum of ego catering and in accordance with the routine work procedures of the "mediocre".
Are you implying that we need front men? That the brilliant prose to date has been written exclusively by distinguished credentials slumming here with the commoners? If brilliant prose occasionally results from the commoners, then our procedure of keeping the best and deleting the rest will eventually result in more brilliant prose being posted. Even with the diploma mills available in the U.S. there are a lot more commoners than P'hds around.
This sort of reminds me of perfect games and bowling. Are the most perfect games bowled by professionals or amateur bowlers?
Lest you think yourself insulted, let me offer an example of mediocrity: my many philosophy articles. They are full of content, they are basically correct, many of them (those that have been re-edited from lecture form) are reasonably well-written--but they are woefully inadequate and basically mediocre. I would be ashamed to bill them as anything other than what they are--very rough first drafts based on lectures to OSU undergrads, which sit there waiting for some experts to, probably, completely rework them, or even replace them.
But no expert will want to do that until the whole project is led by similar experts and therefore, to their mind, there is some guarantee that the project will not wind up being an enormous waste of time. Without that sort of leadership, I fear that my articles, and the many other fair-to-middling (but basically correct and perfectly contentful) Wikipedia articles, will never receive the vetting from qualified people that they really need.
I think you are confuse confusing the term leadership with effective participation. As the Wikipedia content improves in breadth, depth, and reliability; it will attract greater attention from both eggs and chickens.
The guarantee is in the FDL and GPL. My time here is not wasted even if this community collapses or fails to grow. The simple addition of SVG will make the software suitable as a starting point for engineering projects. Indeed, a fellow at advogato mentioned that he took my advice and setup a wiki for engineering specification development. His engineering team is happy with the results.
Several credentialed specialists here have started work on a science textbook or tutorial project. Lead by April, there seem to be heavy contributions from Rgamble and Magnus. Perhaps others have participated as well.
As the project technology matures and the content and community grow in quantity and quality any experts with credentials (with any I.Q. or reasoning ability at all) will be able to see the merit. Until then I propose we struggle along with the precursors, our current participants that have spotted the potential and wish to help develop it.
There is no "leadership" without participation. Perhaps Nupedia has too many Chiefs and not enough Indians. With participation leadership shows up eventually. This is called a band wagon. Leaders must find these band wagons and stay in front of them or merely be participators.
(I acknowledge that an appropriate response to this is: "I agree, but what are you bothering Wikipedia-L about it for? Go post to Nupedia-L." Basically, Wikipedia is the only game left in town as far as the free encyclopedia movement is concerned. If enough of you get behind this, something might happen. To my mind, Wikipedia shouldn't change but Nupedia can and should, and Wikipedia might benefit directly.)
How do you figure only game in town? Nupedia has credentialed editors and participators listed. The Spanish fork has contributors. The Fact Factory appears to have problems but perhaps they are working on that. It is clear evidence somebody is attempting to use the technology and database. Wikipedia's volunteer community and content seem to be continously growing although we are having some growing pains. Much of it seems to revolve around how to mediate conflict, reach agreement that banning thresholds have been exceeded, what are the operational factors governing the community, revenue for the non profit, etc. Any of that sound familar? Perhaps less time shouting "troll" or "liar" and more time discussing issues apparently of interest to at least some of the community members (or former members) would be a wise investment.
How are "trolls" and banning handled at Nupedia?
Get behind Nupedia how? We seem to have plenty of effort required in resolving Wikipedia's current issues. In fact, most of our contributors seem to prefer peaceful editing to discussion of percieved or alleged "ontological" or procedural issues.
You seem to be arguing that we (Wikipedia) need additional leadership from "distinguished credentials" to attract participation from "distinguished credentials". Since Nupedia has plenty of credentials, from my initial brief overview of the site it seemed on average much more credentialed than Wikipedia; what is it you want or need from Wikipedians that you cannot attract from academia via the existing credentialed participants at Nupedia?
If you need grunt work (participation) the FDL license guarantees that Wikipedia's content can be displaced to Nupedia and improved by the resident credentials.
I find your argument contradictory and confusing.
I also fail to understand what you think the bar to participation at Wikipedia is for credentialed volunteers. Surely they will not be ostracised by peer journals or peers for occasionally editing or helping out here? There is also the "anonymous" option unless somebody threatens to "out" them.
Indeed, the peer review experience they bring with them will be quite valuable. Perhaps we can setup a tutorial, some guidelines, or procedures to assist them in conveying this expertise to our neophytes. This does not seem to go over well ganging up on them in the stacks, they merely get defensive and eventually leave one way or another. Occasionally some of us neophytes have problems with engaging in dialogue, discussion, or comment pertinent to the argument rather than the arguer.
Michael R. Irwin B.S. Engineering Physics, OSU, 1984. GPA 3.22 Valedictorian, Coquille H.S., 1979, GPA 3.98 (B in freshmen P.E.) alleged troll and liar proud remnant of the green space faction