The idea of building a centralized consensual knowledge database for the Semantic Web

As it has been written on the main page, computers will understand the texts in the Semantic Web and thanks to it they will be able to answer intelligently to questions asked by Internet users. It is widely assumed that in order to build the Semantic Web we must tag information on web pages to be clear for computers. Then computers would search such a web, collect information and answer to questions.

Below another way of building the Semantic Web will be presented. Instead of tagging information on web pages we could build a centralized database of commonsense and specialist knowledge that could be used by both computers and people to answer on questions immediately without additional processing.

Such a database can be described by analogy to Wikipedia. Wikipedia has enabled millions of Internet users to edit articles collectively. Similarly many people will be able to edit the content of this database. The main difference, but not the only one, is that we would edit individual pieces of information instead of articles. Thanks to the change of size of elementary editing unit from article to simple sentence or object property value, the database would have the following advantages of Semantic Web that Wikipedia lacks:

possibility to get answers on asked questions immediately (for example within a few seconds). Writing simply, it would be possible since it would be sufficient to find in the database the object and its property we are interested in and read its value to get answers on many questions that can be easily reduced to this knowledge.
very easy and fast sharing the information we have. If we want to add some information to Wikipedia then we must add it to some article in such a way that it will constitute an integrated whole with the currently existing information. One cannot add a single piece of information to Wikipedia without writing an entire article about a certain concept. In the database it will be sufficient to fill the value of already existing object property or write a single simple sentence.

Additionally the database would have the following advantages that the Semantic Web being designed does not have:

Confidence that information in the database is correct. It would be reached thanks to:
- evaluating credibility of information and users using an algorithm analogous to HITS algorithm used to evaluate popularity of web pages. Information credibility will be evaluated on the basis of current credibility of users voting for or against this information (by adding an information user also votes for it): the more users voting for information correctness and the less users voting against it, the greater credibility information will have. User's credibility will be evaluated on the basis of current credibility of information he voted for or against: the more pieces of correct information and the less pieces of incorrect information that he voted for, the greater credibility he will have (analogically for voting against the information). Thanks to it moderators will not be needed.
- appropriate verification of user's identity. Thanks to it, among other things, there will be no vandalism.
- obligation for a matter-of-fact discussion with arguments in case users do not agree about information correctness
Users will be able to reach consensus owing to obligation for a matter-of-fact discussion. One can doubt that people would agree about information correctness. However there are many facts that everyone agree with, for example that Paris is in France. Furthermore the scientists' community shows us that thanks to proper familiarizing with a subject and thanks to discussion they are able to reach consensus about more complex topics in most cases.
Simple scalability

Owing to collective editing and reaching consensus the database would be an example of computer implementation of so-called collective intelligence and consensus theory of truth.

It seems that it will be most easy to start the process of building the database by adding to it knowledge about products available on the market. Firstly, products have strictly specified characteristics and secondly building such a database will be economically justified. This application is described here.

Documents:

Description of an idea of building a centralized consensual knowledge database with description of its features and applications
Appendix - comparison of the idea with similar ideas from other fields (philosophy, management, politics) and with currently existing computer systems (available also in Polish)
Description of an idea on Artificial Intelligence forum

Note: In some of attached documents I call the described knowledge database as ontology. However it should be rather called as "semantic net" since when talking about ontology we most often think that it is written using a formal language, can be automatically processed by a computer and used for example for deduction. The proposed system is designed to be used by people not necessarily by computers.