‘Good luck with your WikiPAIDia’: Reflections on the 2002 Fork of the Spanish Wikipedia

Interview between Edgar Enyedy and Nathaniel Tkacz, January 2011

In early 2002, Wikipedia had little more than 20,000 total articles. The project was still overseen by Larry Sanger. It wasn’t yet clear that Wikipedia’s ancestor and first effort by Jimmy Wales and Sanger to create a free online encyclopaedia, Nupedia, would soon be irrelevant. There was no Wikimedia Foundation, no board of directors, no admins or sysops and no arbitration committee. There was no Essjay controversy, no regular media attention and no “sock puppets”. There wasn’t an army of bots working away 24/7, cleaning, ordering, scraping, prompting and reverting the activities of fallible humans. There were barely any “protected articles”. People had to check articles that might attract unwanted attention manually.
The term “wiki” was totally obscure to anyone who hadn’t spent time in Hawaii, but people were still talking about “virtual reality”. Wikipedia still had a dot-com domain, which was owned – along with the hardware – by Wales’ company Bomis. For people who care about technical details, the software underpinning Wikipedia was UseModWiki, written in Perl. Wikipedia’s logo was already sphere-shaped, but the sphere was wrapped with a quote from Thomas Hobbes instead of the now familiar jigsaw design. The logo, along with 90% of the overall project, was in English. The project had begun to internationalise, but exactly what that meant was up for grabs.

In early 2002, the kind of stability that makes it difficult to see the contingency of things had not settled on Wikipedia. People still had very different ideas about what Wikipedia was and what it might become. Sometimes these competing visions produced conflicts, which, like Wikipedia itself, manifest in ways not reducible to historical precedent.

Edgar Enyedy was involved in the Spanish Wikipedia from its launch on May 20, 2001, until mid-February 2002, when he abruptly left the project. Together with the rest of the Spanish Wikipedia community, they took the content they had written to another server, gave it a different name and carried on in a different direction. This reproduction and repurposing is made possible by the copyleft or “permissions based” license attached to all Wikipedia articles. In Free and Open Source Software cultures, what Edgar and the early Spanish Wikipedians did is known as a “fork”. The following interview with Edgar brings this 2002 fork back to life. The purpose is not such much to settle old scores (although there is a bit of that), but to give detail to what we will see is a profound moment in the history of Wikipedia. While interviewing Edgar I also wanted to build a better understanding of the unique nature of conflict in so-called open projects and the related political techniques that respond to such conflicts. What follows is the first detailed, first-hand account of the process of “post-software” forking; that is, forking outside of purely software-based projects.

Edgar was born in Oxfordshire, England, and raised in several countries. His formal training is in Philology and Computer Science and he holds a Master’s degree in Communications Systems and Networking (Polytechnic University of Madrid). He has worked as a journalist, editor, researcher and teacher. He has published in the areas of statistics and social science. He has spent a lot of time working on issues related to networking protocols and has a long history of involvement with the Internet, dating back to “the old Usenet days” (his words). Besides some community-based projects, Edgar is currently steering clear of public life, living in a very small town by the seaside.

NT: Perhaps we should begin with some basic background information. How did you come to be involved in Wikipedia?

EE: Back then, I was studying for a Masters degree in Communications Systems and Networking and I needed to structure and display the info I was handling and gathering in a horizontal network with easy hyperlinking. I tried several wikis and finally I chose UseModWiki, as the programming language in which it was written, Perl, is not that difficult. I checked some implementations of UseModWiki, which first lead me to MeatballWiki, and finally to Wikipedia. Wikipedia was very small. There was a bunch of people claiming that those blank pages would some day turn into an encyclopaedia. Not like Encarta or Britannica, which were our references at that time and both pay-per-consult, but a free one. I started editing, mainly focusing on Talk Pages, as I found errors or incomplete information. I used to come back to those pages, sometimes I left a comment, or maybe I didn’t check back for a week or so. The international projects were just beginning and it soon occurred to me that the Spanish Wikipedia should be the second main encyclopaedia, based on the fact that the Spanish-speaking population around the world was estimated to be over four hundred million (I didn’t think it would be Mandarin, due to the many dialects in China). That’s how I came to collaborate on the Spanish version of Wikipedia.

NT: How active were you on the Spanish Wikipedia in those first six months? How many of you were there? Did you know each other?

EE: There were about 20-25 regular collaborators who worked everyday, editing, reverting vandalism, watching articles and writing new ones. On top of that, there were 30 or so more who visited once or twice a week, but also worked hard to contribute to the project.

Apart from the typical contributions, my role was to communicate with the emerging international community. I was living in Madrid and most other collaborators were not from there. I didn’t go to great lengths to establish friendships, but some collaborators, both from Wikipedia and the Enciclopedia Libre Universal (EL – the fork) have reunited a few times.

NT: There were a lot of open questions about how the emerging encyclopaedias would relate to each other and in particular the English language original, including exactly how they would differ and where they would overlap. How did that play out with the Spanish Wikipedia?

EE: Even when the basic design was set up, there was still an obvious English presence on the Spanish Wikipedia. You might have found Spanish pages in both Spanish and English, even in the same paragraph or sentence. The software, for example, was not translated at all and it cast an English (language) shadow over the entire project. The basic pages (‘what Wikipedia is not’, ‘be bold’, ‘how to start’, ‘sandbox’, etc) were all in English; we had the American logo in English and so on. All we had was an index page and some articles translated or summarised from the American Wikipedia.

This American shadow marked the first point of contention between myself and Sanger and Wales. Since they began from scratch, I thought we should do just the same. The Spanish encyclopaedia could not be a mere translation of the English Wikipedia. The organisation of topics, for example, is not the same across languages, cultures and education systems. There are also quite different perspectives regarding censorship. Former AOL users used to remind me that explicit biology images are widely accepted among us, but would be considered inappropriate on the American version. Historiography is also obviously not the same. We are used to our own history schemes and the American one didn’t fit at all. Basically, it became very clear that the American template would not fit the Spanish project.

At that time, all the Wikipedias had an index on their first page and that index seemed entirely strange to us. I worked hard on creating a new one, dealing privately with Wales over email and publicly with Sanger on the mail list. I worked from eight to twelve hours a day for six months to get the Spanish Wikipedia working and to make it more attractive for users. We even set up an alternative index based on the Universal Decimal Classification, with templates for biographies, geography and so on. From the HomePage you could switch to that index if you felt more comfortable working that way.

I also started to develop a “Wikipedia Style Book” for the Spanish language version that advised on how to deal with acronyms, long and compound surnames, the use of bold and italics and so on. Our editing policies and rules were very similar – we were all Wikipedia – but not the content or classification method. This Style Book came from my background in journalism. It was warmly welcomed by the community and was widely used. At the time, the idea was not adopted by the other Wikipedias.

NT: What about the relationship between the Spanish and English language communities during this period?

EE: The relationship was a strange kind of tolerance from the American staff. They knew for sure that they couldn’t afford to let us go, as each and every international project was receiving a lot of attention. The international wiki list was watched carefully, not only by the international community, but also the American community. They paid close attention to how things were developing.

NT: You are already hinting towards the fork, but first I want to get a sense of Larry Sanger’s early role. From the early discussion lists (archived on osdir.com), it seems like Sanger very much acted like the leader or at least “facilitator” of the entire Wikipedia project. Is this how he was generally received by the Spanish Wikipedia?

EE: Larry Sanger acted as a Big Brother. He was an employee, a Bomis-Wales wage-earning worker. I can’t stress this enough. Nupedia’s failure left him spare time and he was allocated by Wales to Wikipedia. I really regard him as a co-founder of Wikipedia, even though this fact has become less visible over the years. There were two people heading the project, and it was difficult to tell where the ideas came from.

The American Wikipedia might have seen him as a “facilitator”, but we regarded Sanger more like an obstacle. At that time he was not an open-minded person. I have to admit that he brought some good ideas to us, but the American Wikipedia was too caught up in the interests of Bomis Inc.

I engaged in head-on confrontations, open clashes, with Sanger. We were all working on a basis of collective creation, with peer-to-peer review. It was an open project, free in both senses. We were all equals, a horizontal network creating knowledge through individual effort – this is the most important thing to keep in mind. But Sanger turned out to be vertically minded. His very status as a paid employee led him to watch us from above, just waiting for the right moment to participate in active discussions in the (mis)belief his words would be more important than ours.

NT: The most significant of these open clashes, the one that lead to your departure from Wikipedia, was sparked by a seemingly insignificant remark, made by Sanger in passing, about the possibility of incorporating advertising in order to fund his future work on the encyclopaedia(s). His exact words were ‘Bomis might well start selling ads on Wikipedia sometime within the next few months’. From your reply, it was very clear that you were against ads, but more than that, it seems like this was a decisive moment, the straw that broke the camel’s back, as they say. Can you revisit this event and tell us how it unfolded?

EE: The possibility of advertising was out of the question. I asked Wales for a public commitment that there would be no advertising. This only came after we left. There were, however, other things that I was not happy with, some pretty straightforward, others a little more complicated:

• All Wikipedia domains (.com, .org, .net) were owned by Wales. I asked myself ‘why are we working for a dot-com?’ I asked for Wikipedia to be changed to a dot-org.

• I wanted the Big Brother out. Larry Sanger was against the nature of the project itself. None of us felt comfortable with such a figure.

• I had asked for the autonomy of each foreign Wikipedia. We did not want to be seen as mere translations of the American version. We asked for things like our own logo, and Wales agreed, but it was clear that he didn’t consider the international wikis as an addition to the ‘main wiki’ – all the best articles were there, as well as the most contributors and total articles. I was told so many times to translate from the main wiki, and my response was always the same: We are not a translation of the American Wikipedia!

• There were significant software issues. The latest software releases and revisions were only installed and running on the American Wikipedia. The Polish Wikipedia, for example, could hardly develop at that time due to problems dealing with special Polish characters. All of the international Wikipedias were running out-of-date software and because Bomis Inc. controlled the wiki farm, we couldn’t do anything about it. I asked for access to the farm (just the Spanish server), but after a short discussion my request was denied. They said it was for security reasons because Bomis Inc. was hosting files from its clients on the same server. As we couldn’t access the wiki farm, I asked for mirror servers to be set up over and over again. The answer was always the same: that we needed to keep the project together. Wales added that there were some technical reasons for why they couldn’t set up a mirror site, but he couldn’t explain what they were (and didn’t even seem to believe them).

• Wales had stated his future intentions of making hard copies from the encyclopaedia(s), noting that it was permitted under the GNU/FDL license. It clearly was part of the license and I agreed with the idea. I told him, however, that the organisation that initiated such a project would necessarily be a foundation, and not just one, but rather a foundation in each and every country. I saw the project as completely non-profit and thought our goal shouldn’t be to figure out how to pay wages. Wales always replied that a foundation was very difficult to set up. I told him it was an easy deal: you are contributing to the project with the servers, we are giving our time and effort in an altruistic way, but no one is going to make money from the project unless it is proven that the money goes to people who really need it – and that doesn’t include staff members.

• When I asked Wales through private emails to set up something – to set up the Basque Wikipedia, for example – he always replied: ‘I’m not a wealthy man’. I heard that many times. A couple of years back he said in an interview ‘I don’t care about money’[1]. When I think about this position and those exchanges, it makes me laugh. Wikipedia has created a large foundation of wage earners, and each year he has to ask for ever-increasing amounts of money. This is what I didn’t want to happen: a large, money-centred organisation made possible by the free work of the community. After we forked, he wrote to me and said: ‘There will be a foundation and a place for you is waiting there’. It was clearly an implicit deal: You all come back to our project and our servers, and I’ll reward you. The fact is that I wasn’t looking for a seat on a foundation. I just wanted the whole project to work the best way we could (or knew how to).

Because of these things, I didn’t trust Wales’ intentions. Not at all. We were all working for free in a dot-com with no access to the servers, no mirrors, no software updates, no downloadable database, and no way to set up the wiki itself. We were basically working for Bomis Inc., and asked in a gentle way to translate from the main Wikipedia. Finally, came the possibility of incorporating advertising, so we left. It couldn’t be any other way.

I would like to remark upon the fact that as it is known today, the international Wikipedia that you all know and have come to take for granted, might have been impossible without the Spanish fork. Wales was worried that other foreign communities would follow our fork. He learnt from us what to do and what not to do. The guidelines were clear: update the database; make the software easily available on Sourceforge; no advertising at all; set up a foundation with a dot-org domain and workers chosen from the community; no more Sanger-like figures, as well as some minor things I haven’t mentioned, such as free (non proprietary) formats for images.

NT: During the discussions about leaving and forking, you were very active, but you also note that others shared your opinions. Were you leading the revolt (as it is written on the EL entry on the English Wikipedia), or were there other influential/respected people with significant roles?

EE: You could say that I was some sort of unofficial leader together with Javier de la Cueva, and yes, others shared our opinions. Sadly, there weren’t other influential and respected people with significant roles. Many remained anonymous. I did, however, receive a lot of support from the community. Some offered money, others offered help with hosting and securing a domain. It was Juan Antonio Ruiz Rivas who organized hosting with the University of Seville, as that is where he worked.

I recognised that people wanted to make suggestions, to debate and be heard. But those kinds of processes can be lengthy, so I made the decisions. I thought the timing was critical – a line had been crossed and I didn’t want it to be a never-ending story. Luckily, the community supported me. This was the extent of the unofficial leadership: I made a decision and others supported it.

NT: In the small body of literature available about forking, it is often assumed that forking is as easy as downloading an album. Although the ‘right to fork’ is thought to be an essential aspect of open projects, the actual details of forking are rarely considered. What exactly happened when you decided to fork? What were the decisions that you were faced with (regarding content for example)? Did it require much technical expertise?

EE: At that time, to set up a wiki and to export the .tar database from Wikipedia was almost impossible. The GNU/FDL license granted it could be done, made it legally possible. But no way! The Wikipedia page on Sourceforge had instructions that read like hieroglyphics. And once again due to ‘technical’ reasons (that none of us believed), the downloadable database was never updated. I asked Wales about the wiki itself and the database and he just replied ‘in the future’. It was not fair. These conditions did not resemble what the GNU/FDL was supposed to ensure.

I remember after I wrote ‘Good luck with your wikiPAIDia’, I started receiving messages like: And now? What’s next? The first thing I thought about was looking for a hosting company and registering a domain. I was also thinking about how we could make this component effectively community-owned. I had the idea, for example, that we could change the domain registrar each year so there was not a single continuing owner. There were few hosting companies with the characteristics I was looking for. Remember, at that time, to work on the server side was not as usual as it is today. In actual fact, one of them was Bomis, but hosting with them would be a cruel joke. Javier de la Cueva, who is a very well known lawyer, offered his domain as well, but as mentioned, we ended up getting hosting from the University of Seville.

Setting up the new encyclopaedia wasn’t an easy job. I began by configuring a spare PC as an Apache server and started working on the software. The Perl scripts ran OK and the wiki could be reached through a proxy server from other computers on the net. ‘Well’, I thought, ‘it runs’. It took me a week to get it going, but this seemed a very small amount of time when compared to the dozens of hours I spent arguing about the project with Wales and the community. The Spanish community had worked very hard on Wikipedia. I remember writing a lot of articles on Computer Sciences and Literature, making Indexes, developing subjects and so on, and the rest of the community was just as active. When the server was up and running, and as the GNU/FDL permitted, we began copying our articles from Wikipedia. Is wasn’t an automated process, no bots or anything, just us bringing the articles across one by one from Wikipedia’s server to ours. That was the beginning of EL and it was the strongest time for the community. I also started sending individual emails to hundreds of town councils and tourism offices, asking them to participate. About 10% joined in, writing pages on their respective towns, which was a pretty good response rate.

Our actions made Wales realize how the whole project could be hosted on non-profit servers all over the world. Others could follow in our path, so he had to change things quickly on the American and international Wikipedias.

NT: Once the fork – titled the Enciclopedia Libre Universal (EL) – was set up, how did it differ from the Spanish Wikipedia?

EE: We had realised that a lot of content on the Internet was the same, maybe slightly changed, but practically the same info across different sites on a chosen topic. If you wanted to find out about a particular museum, for example, the info you received from Wikipedia was just the same as you would get on the official page of the museum itself, slightly converted, and reworked, like (bad) school homework. We wanted quality over quantity, and original articles, not carbon copy.
This is one of the many things I criticise today: Wikipedia has led us to a verbatim information Internet. There used to be a lot of different sources, but nowadays the info you get is carbon copy all over the net. There aren’t enough filters. A lot of pages are just circulating Wikipedia texts, including its rights and wrongs, but without its disclaimers.

I had also suggested that we begin some articles only with links, or just a small stub with links. There was already some very high quality information about many topics, both from official and non-official pages and sources, and there was no sense in reworking all that material. Just an article with an official link would suffice. I was told that this was not the ‘proper way’, as they (Wales and Sanger) didn’t want to look like Dmoz. Of course, today Wikipedia pages are full of links to other sites.

NT: While the Spanish Wikipedia stalled severely for at least a year after the fork, after two years it had bounced back and was already larger than the EL. Today, the Spanish Wikipedia has almost 700,000 articles, while the EL has more or less flat-lined at around 45,000 articles. Is there still a community around the EL? Did anyone go back to Wikipedia?

EE: Nowadays, almost all EL members belong to Wikipedia too. There is still a working community. However, it is wrong to think (as Wales had) that EL contributors are duplicating the work they do simply because the CC license allows the content to be transferred to Wikipedia. The truth is that they enjoy working without Wikipedia’s guidelines and structure above them. They choose their own policies. A lot of the time EL contributors would upload their own articles to Wikipedia, but that wasn’t necessarily the main goal.

NT: While it would be easy to look at the numbers and conclude that in the long run the EL failed, I think it is clear that the fork had a significant impact on the direction of the entire Wikipedia project. As you have stated, after the Spanish editors left, Wikipedia decided not to have ads; it changed its domain to dot-org; it upgraded a lot of the software; and it set up the Foundation to oversee the project.

EE: Right. The fork had its time and place, its goal and its consequences. Nowadays, the romantic point of view is that EL survived and is still going strong. It is a nice view, but wrong. EL has failed as a long-term project for one reason: The project itself was not intended to last. It was merely a form of pressure. Some of the goals were achieved, not all of them, but it was worth the cost.

NT: For a while there was talk of officially reuniting the projects, but it never happened. What was the relationship between the encyclopaedias after the fork?

EE: Both encyclopaedias linked to each other, and shared contributors. A lot of valuable people left Wikipedia. But there’s a life cycle for collaborators and newcomers reached Wikipedia first. The reunion never happened because EL wanted to protect and preserve the free space it had carved out for itself – some sort of oasis. Nowadays I would like to see them back on Wikipedia, working on the same project, reunited at last, as the EL mission is accomplished.

NT: What do you think of Wikipedia today?

EE: Today, Wikipedia has become a huge, hierarchical social network, behind an unreliable knowledge repository. That’s what it is, merely an unreliable repository. As the project continues to grow, so does Wales’ celebrity status, but the same cannot be said about the quality of the project, which is being left behind. Wikipedia has reduced the minimal requirements of knowledge to below average in both quality and reliability.

The rise of fundraising campaigns also shows what Wikipedia is not: free. During the 2010 campaign, Wikipedia received $16 million in donations. It is often said that Wikipedia competes with the Googles and Facebooks of the net on a fraction of the budget, but Wikipedia never had to play this game at all. If anything, the foundation should be generating revenue, though not through selling ad space (the original idea was to sell hard copies). As we speak, the foundation is also offering scholarships to attend the annual ‘Wikimania’ event. All revenue should go towards realising Wikipedia’s main vision of distributing knowledge to those who need it most – this certainly doesn’t include providing scholarships to its own events.

NT: Would you do anything to change Wikipedia?

EE: Wikipedia is working well the way it is. It is what Wikipedians want it to be. There are a lot of people involved in carrying on the project and this is what they have chosen. It’s not my kind of project, not my social network, so I’m not a user. I dislike Facebook, Twitter and Wikipedia policies, so I stay away from them. There is a lot of work to be done to change Wikipedia, and I guess I am in a minority.

[1] Edgar is referring to a comment made by Wales in the Catalonian newspaper La Vanguardia, January 8th, 2009.