- From: Joshua Allen <joshuaa@microsoft.com>
- Date: Wed, 31 Jan 2001 20:11:36 -0800
- To: "'Aaron Swartz'" <aswartz@swartzfam.com>, www-rdf-interest@w3.org
Aaron, thank you for your comments. Some response below: > The benefit of Nupedia is that it is free (libre). Thus people are allowed > to modify, translate, copy and sell Nupedia articles without having to > worry about copyrights. Also I think that another motive was to eliminate the bias that corporations might selfishly impose. Even Britannica I suppose was an instrument of British cultural imperialism. Nupedia will probably be biased towards the interests of people who are opposed to copyrights, right? > (swinging back around to RDF) Interestingly, the project has been > ways of cataloging articles without forcing a cataloging system upon the Yeah, that's what I meant by them trying to build the semantic web from scratch. First, nupedia allows anyone to contribute content, so no different than WWW. Next, nupedia gives you the ability to look at only those articles you "trust", by artificially creating a community of people with similar interests (anti-copyright). But the point of the semantic web should be *inclusive*, IMO. That is, anyone can publish metadata, and you should be able to choose to "trust" content based on any definition of community that you wish. Just like you should be able to categorize in different ways. Here is a scenario to describe what I mean: 1. You are browsing the web, and you see a page that you think is good. You have a thumbs-up and thumbs-down toolbar buttons in your browser toolbar. You click "thumbs-up", and a small XML packet containing your e-mail address, the URL in question, and your metadata gets silently sent somewhere. 2. Somewhere is a server containing your information and a set of groups to which you belong. Membership in the groups could be decided in a manner similar to advogato, free to all, or however the person creating the group wished. Anyone could create groups. You could use some UI at this server to rank any groups you were interested in and weight how much you trust or distrust each group's metadata. 3. Periodically, maybe every hour or so, each metadata collection server (where you sent the xml packet) aggregates the metadata based on how many of each metadata value were sent by members of each group. The aggregates are sent to various services that have subscribed to the metadata (google and hotbot, for example). 4. You do a google search on certain keywords, and the results are automatically ranked to show those results that you would find the most trustworthy, and filter out those that you did not trust, all without ever having to modify the original pages. [Note that nothing about this example claims that the "metadata collection servers" have to _only_ collect "sucks/rules" info about pages, or that subscribers have to subscribe to _all_ of the metadata or have it aggregated by groups, or that the "metadata servers" have to be centralized, controlled by one organization, etc.] IMO, this is one of the major underlying visions of the SW.
Received on Wednesday, 31 January 2001 23:23:54 UTC