Jump to content

User:Syced/Semantic Wikipedia

From Wikipedia, the free encyclopedia

Goal

[edit]

Express the world's knowledge as "triples". A "triple" is an invidual element of information that a computer can understand, in the form subject-verb-complement, such as Mozart-died-1791. Each triple must have at least one reliable source.

Grammar

[edit]

The exact format is not settled, RDF seems to be a safe choice but other formats might be more readable. For now, let's use this simple format:

  • Subject
  • Verb
  • Complement
  • References

Languages

[edit]

Semantic information can be understood by machines, but it can also be translated into human-readable information. Thus, even a very small language could have a potentially huge encyclopedia, after just translating a few templates. The generated sentences will not be as stylish as if they had been written by Shakespeare or Goethe, but still perfectly informative.

Editing raw triples is not fun everyday, so a software should be created and used to add/edit information without manipulating the technical details. This software could be localized in many languages, so that people who only speak Swahili (for instance) can edit as well. The English Wikipedia has a lot of contributors because a lot of people can speak English, but this Semantic Wikipedias could have even more contributors, because people would not need to know English to contribute.

This could become the first truly "universal" encyclopedia, which would not be biased by the cultural background that surrounds any particular language. Speakers of any language would be able to contribute. Thanks to a well-defined and objective grammar, few point-of-view conflicts should arise. Moreover, no information could be entered without citing a reliable source for it. A system (yet to be defined) will make sure at least one "confirmed" editor who understand the source's language review it to make sure it is reliable.

Draft

[edit]

The Semantic Wikipedia must start somewhere, so here is a draft of what could be the first few triples:

(Subject1 is a person)

(Subject1's name is James Gosling)

(Subject2 is a book)

(Subject2 has Subject1 as an author)