Ruth Ifcher cannot remember exactly when she first learned of Nupedia, but one thing she does remember—it was Sanger’s infectious enthusiasm that hooked her on the project. A computer programmer by day, and a former copy editor and holder of several higher degrees, Ifcher was someone Sanger depended on in his early editorial team.
While Sanger as editor in chief was the sole paid position in Nupedia’s ranks, Ifcher agreed to volunteeer as chief copy editor. Even though she was located on the other side of the country in New York, she and Sanger worked closely on Nupedia’s early policies and working procedures. She was one of the early ones who signed up for high positions in the newly formed Nupedia Advisory Board to provide direction.
Sanger was given broad authority to set up Nupedia’s working process and imagined a tiered structure, combining his vision of an academic quality process with tapping the use of volunteers on the Internet. Among the roles he proposed were writers, editors, and copy editors. In Sanger’s vision, he still believed in a top-down structure to manage quality:
Editors assign topics to writers and formulate any necessary policy and direction with respect to their own categories, though they are expected to follow and enforce general Nupedia policy guidelines.10 (It must be noted: This is quite different from the concept of a generic "editor" in Wikipedia, which refers to any individual, even an anonymous one, who simply modifies a page. When talking about Nupedia, "editors" are ones who have been vetted and have extra authority over mere "writers.")
His academic roots compelled Sanger to insist on one rigid requirement for his editors: a pedigree. "We wish editors to be true experts in their fields and (with few exceptions) possess Ph.D.s.," read the Nupedia policy. Editors, in Nupedia’s parlance, were "in charge of particular subject matters."
Editors meeting the Ph.D. requirement still had to have some formal way of proving their status, usually by reputation in academic circles or publishing record. Sometimes it wasn’t so easy. Building an "academically respectable" encyclopedia would have to rely on bona fides, but when information was lacking, other means were needed. In one case, where the person in question had no Internet Web page as proof, he was asked to fax a copy of his degree to Sanger.
This was something Sanger wanted to emphasize. While the authors were drawn from anyone who wanted to apply, they were subject to review and approval:
The category editor then decides whether to assign you the article topic (step one). The editor may, in some cases, ask what qualifications you have to write on the article, and therefore it would be a good idea to have completed your member profile. (This is private information.)11
The first thing the Nupedia Advisory Board did, by the summer of 2000, after email discussion, was to settle upon a seven-step process for all articles. This consisted of:
1. Assignment
2. Finding a lead reviewer
3. Lead review
4. Open review
5. Lead copyediting
6. Open copyediting
7. Final approval and markup
Before the Nupedia process even started, volunteers were grouped into their subject areas. Three or more people were necessary to make an "active" subject.
Of the dozen or so subject areas active at any one time, biology became known as the deepest and strongest in Nupedia, because of the preponderance of scientists and academics on the Internet.
For an article to start to run the gauntlet, someone first had to create a proposal with a description of one hundred words or less. After review, editors in the proper subject area would hopefully approve the proposal for assignment. The author would then draft the article, whereupon a volunteer editor acting as lead reviewer would review the content and accuracy.
The lead reviewer was responsible for getting the basic information to a level the reviewer was happy with, or in extreme cases the article could be rejected altogether. Correspondence between the reviewer and the author was mostly done by email, though some parts were done via a Web site. Members could not view the articles in the peer review process. Until the material was approved, it was a closed loop between the reviewer and the author.
When the lead reviewer was satisfied, the article was then brought to open peer review, where every member of the community was allowed to comment on it and appropriate modifications would be made.
The entire process to this point would take roughly two weeks, before Ifcher, as lead copy editor, would get her hands on the article. If it got this far in the process, it was probably going to go through all the way. "Very few things were aborted in the middle," she recalls.
With the subject’s content and accuracy approved, Ifcher would assign two people to the copyediting task. One would do the heavy lifting of the major copy-edit, to correct footnotes, grammar, punctuation, and style. The second would look over the work of the first. The procedure specifically called for finding folks with a fresh angle on the article’s subject.
Concerning copy editor selection, Ifcher wanted diversity: "You were encouraged to pick articles in fields you were not familiar with, on purpose, to make sure the article was clear." The Nupedians made it a goal to create articles a college student could understand if the student had no background in the subject.
The final community step was the open copyedit, when everyone in the community could make final comments and propose fixes. This was the absolute last chance for anyone to raise problems or objections. In general, the copy editor, rather than the author, would answer comments and in the best-case scenario, every change would then be cleared and approved by the author.
The final stage overall was for formal mark-up by the editorial staff to make the article of publishing quality, containing the right headings, spacing, and so on. Sanger would have the final approval and, at last, a Nupedia article would be born.
In September 2000, the first article to completely make it through this process was called "atonality" by author Christoph Hust, a German music scholar of the Institut der Johannes Gutenberg-Universitat. It read:
ATONALITY BY CHRISTOPH HUST
Atonality means the absence of a tonal center as the basis for the whole or part of a piece of music. This was one of the radical innovations in music around the second decade of the twentieth century. More specifically, the term is used to describe works that are neither tonally-centered nor use the 12-tone or dodecaphonic [doh"-dek-&-fon'-ik] method of organizing pitch. Although atonality is primarily associated with the composers of the second Viennese School, namely Alban Berg, Arnold Schoenberg [shoen'-beRg], and Anton Webern [vay'-beRn], it was also used by other composers such as the Austrian Josef Matthias [yoh'-zef maa"-tee'-aas] Hauer, the American Charles Edward Ives, the Soviet composers Arthur Lourie and Nikolay Roslavets, and the Pole Jozef Koffler. In France, Charles Koechlin’s work approached an atonal style as well.
At the end of the article were fourteen footnotes, twenty-seven references for further reading, and a discography reference of recordings. Telltale signs of the rigorous Nupedia process could be found at the bottom of the article:
Posted 2000-09-11; reviewed and approved by the Music group; editor, G. B. Lane; lead reviewer, Edward (Ted) McIrvine; lead copyeditors, Bruce Hamilton and Ruth Ifcher.
The bar for creating an article was set quite high. It was a system managed by elites that demanded high qualifications and a rigorous fixed procedure. As a result, it created a big bottleneck.
"The problem was it took forever for an article to get through," Ifcher recalls with frustration. "The first year, if we produced twelve articles, I’d be surprised."
Sanger seems to remember about two dozen articles with 150 left in draft stage. But all estimates for that first year were measured in "tens" of finished articles. The process was not scaling. They had a very rigorously engineered solution, and it was not clear how to speed it up.
Sanger tried to move away from the "clunky mailing list system" they used for collaboration and turn to a Web-based solution. In the fall of 2000, Nupedia hired a programmer to write custom computer code to manage the encyclopedia’s complex process online.
But it proved to be no better than the old method. "By the time the web-based system was ready . . . it had become obvious to Jimmy and me that the seven-step editorial process would move too slowly, even when managed on the web," lamented the editor in chief.12
In retrospect, Sanger realized that having both a high bar for contributors and a rigorous process provided a double obstacle. It simply hindered creating the necessary critical mass. "As it turned out, a clear mistake of mine and others was to assume that such a complicated system would be navigated patiently by many volunteers, even if they had clear enough instructions. That is a mistake I doubt anyone designing volunteer content creation systems will make again."
Perhaps the death knell was when one of the founders himself lost faith in the project. Wales wanted to try his hand at what he had funded and took up writing in his area of expertise—he started to pen an article for Nupedia on options-pricing theory, a field he had studied as a Ph.D. student.
The prospect of his work being sent to finance professors for review was the "aha" moment for him. "I had been out of academia for several years. It was intimidating; it felt like homework."13
As the first year was coming to a close, it was clear that the process was not going to scale up in a practical manner. Even a tenfold increase in volunteers would not provide a sustainable return for Bomis’s time and money. Wales and Sanger were searching for solutions. Bomis partner Tim Shell was glad at the very least that "they were looking for ways to speed it up rather than to shut it down."
Nupedia was too much process, too little volunteer output, and not enough money. And it most certainly wasn’t fun.
Something had to change.
