MeatballWiki | RecentChanges | Random Page | Indices | Categories

Does Wikipedia have to fork?

Can't a way be found to resolve the tension between the following values?

It is technically possible for a resource such as Wikipedia to keep both the chaos and integrity, but it requires the revision history to branch, in the fashion of release engineering. You have a CURRENT branch that everyone and their dog can hack on, and a STABLE (or in this case, REVIEWED) branch where peer-review from very select groups takes place before any new revision can be committed. There will be far fewer articles with STABLE branches, and they will evolve slower, but that's the tradeoff. There are more complex alternatives involving simply rating articles among multiple criteria and using a customizable trust metric. Such ratings would be lost at each revision, but presumably your own peer group has interest in reviewing changes quickly. Such a system is likely to devolve however into cliques that merely wish to reinforce their own worldview, actual truth be damned, but cliques are always a problem among self-selecting groups. In any case, Jimmy is highly unlikely to implement such technical measures on Wikipedia for many and various reasons. --ChuckAdams

ChuckAdams said "In any case, Jimmy is highly unlikely to implement such technical measures on Wikipedia for many and various reasons." What exactly are the reasons? I have thought about it carefully, and I've decided that the solution is RightToFork. I'm going to code and host my own engine, importing content from Wikipedia as a starting point.

I think it is a great way to contribute to actively experiment with your ideas and share your results with others in the mainstream project. -- SunirShah

I do not understand SunirShah's comment. I am sufficiently disgusted with the mainstream Wikipedia that I now view it as a read-only resource. I think the situation there has deteriorated to the extent that a clean break and fork is required. I *AM* going to code my own engine and host it myself - it's just a matter of me finding the time, learning mod_perl well enough to do it, and incorporating to protect myself from personal liability. Integrating it with the mainstream project is probably not an option, because my engine design will be completely different. I doubt they would agree to adopt my engine until they lost a substantial number of users to my fork-site, at which point integration would be unnecessary. For example, displaying the IP address of all contributors is standard here, but registered users are completely anonymous on Wikipedia. I don't think I could get Wikipedia to change that policy - display the IP address of all edits. Therefore, I'm not even going to bother arguing other points. Actually, I consider discussing things on Wikipedia to be a waste of time, because of the ridiculously high noise-to-signal ratio.

I have seen other "improve the engine" proposals out there, and have read them and combined the best ideas I've seen into my own specification. However, those seem to be mostly idle speculation. I am at the "about to start coding" step for my own engine. I've selected mod_perl and PostgreSQL? as my language and database for implementation (unless you make better suggestions).

I have been thinking about the "Wikipedia is broken" problem. I've come to the conclusion that the solution is better software - a better engine.

However, I don't view this site as a read-only resouce, so I'm perfectly willing to discuss my ideas here with you. However, I request anonymity (for now). So, please don't give me a hard time for violating the RealNamesPlease policy on this site. Actually, I have my own idea for implementing a RealNamesPlease policy, which I will call ExtremeRealNamesPlease? (which I intend to use on my fork). I'll fill in the details if you're interested.

Well, don't forget that software is made by people, and those people are involved in a political process. Different software just means different politics. If you're disgusted with Wikipedia now, it's better to channel your energy into building your new engine rather than become embittered. Then you'll have something tangible to show when presenting your ideas to others, which should improve their receptiveness. Hopefully by then you'll have enough distance from Wikipedia to be willing to share ideas with them in a warm, positive, friendly manner. -- SunirShah

Your reasoning is somewhat circular. Essentially, you are saying "once you have a working engine and some users, then other people will take you more seriously". However, once I have a working engine and users, why should I care what people on other websites do? I agree that the next step is for me to actually implement it and host it.

Besides, suppose I'm right and my engine really is super-amazing. Why should I give it away? Even though I think software patents are abhorrent, why shouldn't I patent it and host it myself or sell it. (Yes, I disapprove of software patents, but selling it for a nice chunk of change would enable me to do other good things.) Of course, you won't be able to see for yourself until it's done. However, if I do all the implementation work myself, I don't really see why I should feel obligated to give it away.

As a long-time wiki developer I see you head-in-the-clouds. There are at least 50 wiki engines actually actively competing in the "market" and its really hard work to be innovative, get attention and be successful. Most of the engines are GPL-ed, so a majority of users=founders=admins just take a free engine into account. At the moment there is really not a single reason - sorry for this - to trust you to succeed. -- HelmutLeitner

It's somewhat (and perversely) reassuring that other people consider my goal to be hopeless. That way, I don't have to worry about serious competition and can take my time. I have decided on my specification, and it is interesting that you judge it a waste of time without hearing the details. Anyway, if someone demands I open-source my project at a later date (assuming success), I'll just point them to this page and say that I feel no moral obligation to release my code. Yes, I have heard other head-in-the-clouds proposals out there. The difference is that I intend to implement it, and have the (IMHO) ability to do it. Starting from the right tools, writing an engine actually doesn't seem to be that much work. It may take a year or two, but with no serious competition, I guess I don't need to rush. (By competition, I mean that there are competing engines, just no competition from engines with most of the features I'm planning.)

It seems you like the HeroRole?. Knowing better than WikiPedia and everyone, fighting your lonely war. But the facts are much less heroic. To create a new idea, explore it with users, and make it understood and popular is 10x more work than cloning it later without risk. Engine developers copy whatever they see and like, quickly. Next is that you can't build an engine on a small subset of features, because all major engines support 150-200+ features. 300+ features in two years. So be prepared to add another few thousands work hours to be on par with them. -- HelmutLeitner

I concur with HelmutLeitner. Plus, I still do not see that anyone has made a clear case as to why Wikipedai MUST be forked. Many people will feel the same way. Many people will guess that any problems in Wikipedia can likely be solved by working within the group that makes up Wikipedia. It seems logical for me to work with the Wikipedia community to solve problems and issues, rather than just fork it. You've said that you think forking is the only option, but I don't understand why you see it that way.-SamRose

It is a natural process to fork, through RightToFork. WikiPedia is highly successful in many ways, yet there are scores of people interested in collaborating who find it frustrating. LionKimbro and I have discussed this and he calls his ideas for a next step a FederatedWiki? and I call mine WikiVersioning?, both are thinking about the many points of view and many authorities on a given subject. NPOV and "TheOneArticle?" cause a tension that is good at one level and frustrating at another level. GeorgeBush/RepublicanView - GeorgeBush/DemocraticView - GeorgeBush/ChristianConservativeView - GeorgeBush/AnarchistView - GeorgeBush/LibertarianView - why limit? - wouldn't it be good to have a system that worked so people could work on their own individual view of a topic. This could open up a role for someone to try to condense them onto a main article (like Wikipedia does) or GeorgeBush/AllViews - then based on what people trust (sources of the writing) and agree with could get linked to. I say good luck with the forking, but as this community is saying it is not simply a technological exercise, you will need to collaborate too. This has been rattling around in my head for a while, and not as refined as I would like, but it is the gist of it. (see WikiIndex:WikiEngineTree? at http://WikiIndex.com/WikiEngineTree) Best, MarkDilley


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions