Re: [Wikipedia-l] The future of Wikipedia

On Thu, 16 Aug 2001, Krzysztof P. Jasiutowicz wrote: I just have a few random, not-worth keeping comments so will reply here rather than on the wiki site. (Btw, good idea to post it there.) Personally, as long as Bomis is providing tarballs and enough to *potentially* allow for forking, we've zero incentive to do it at all. Sort of, the more control they're willing to give away, the more control we can all trust them to have, I guess. No one in the world expects credibility from Wikipedia. And Wikipedia requires nothing from the world at large to be successful. Thus to me it seems like there is only a tenuous connection - at best - between credibility and success, regarding Wikipedia. If Wikipedia were *only* intended to be a literal replacement for a traditional encyclopedia, sure. But it's something a tad less, and something a tad more. That said, I do not think that the lack of credibility is as clear cut as would appear at first glance. Yes, logic says one should expect to see a distinct lack of credibility in Wikipedia. However, the evidence we're seeing is that many articles actually *are* reliable and credible. In a small but growing number of cases, the articles are actually *better* than you'd find in a traditional encyclopedia. Well I have disagreed with Larry on many things, but on this particular one I think he is correct. I've been involved with Wikipedia since the start, and have watched the evolution of many articles. I think this "self-healing" is not an expression of an idealistic wish of his but a characterization of a real thing that we have been observing again and again. I don't know that I would go so far as to say that self-healing will ensure that at some tangible point wikipedia will be 100% correct. Actually I'm fairly confident that will never happen. But then, does that matter? No encyclopedia is 100% correct, and probably not even 75% correct, when you consider how much is unknown or incorrectly known in the world. I would like to analogize to science here. Scientists 300 years ago did not say, "Let's make sure we have everything completely and reliably figured out as soon as possible, and record it." Instead they came up with a process that allows for establishing what they believed true, with processes for testing and validating and adjusting as we go. A self-correcting approach to accumulating knowledge. With Wikipedia we are using a similar approach - we record what we believe to be true, and then it is challenged and tested, and hopefully replaced with something better. I agree. However, to me it seems like this describes what Nupedia's role has become, and I think it already serves this purpose very well. The wiki cgi is actually pretty light on resource usage. The format of the html pages are very simple and so bandwidth is probably not a major concern. We seem to have weathered the slashdotting extremely well. So distribution for purposes of resource distribution isn't going to buy too much, IMHO. Larry and Jason and etc. can let us know if it becomes a problem and we can seek out a solution then. A second reason for distribution is to make it harder to "kill" the service, ala Napster. But as long as tarballs of the site are available periodically, this is not a major concern. A third reason would be for caching purposes. If a large site like google is the source of many frequent accesses of the site, there might be benefit to establishing read-only caches. Again, though, I think this is Bomis' domain to worry about. If it ain't a problem for them, it ain't a problem for us. Agreed, agreed, agreed. What we need, essentially, is an upload tool that lets us post images onto the wikipedia site. This opens many benefits but also some cans of worms, so this is not a trivial request. I've written file upload systems several times over the years, and I'm about to need to do yet another one at work. I will *tentatively* offer to volunteer to provide something that can be plugged into wikipedia to do file uploads, if Larry and Jason give the go ahead for it. It may be a few months, if at all, before I can have it ready, though. I think this is a good idea. Many pages require editing attention. This is a topic I've heard Larry and others discuss many times. The issue is just finding a way to incentivize folks to do this. I think we are still searching for the solution here. I think we're ok on this one now. usemodwiki is available openly and has been for a long time. (In fact, I'm using usemodwiki on half a dozen other sites right now.) Also, the recent tarball includes the usemodwiki software, and all of the associated scripts and such. Bryce