Database of Wikipedia Will Answer Questions From Internet Users

The Wikimedia Foundation has a new, grand and important project in the square. Initially developed by the office in Germany, the wikidata begins to be developed and future risks to be one of the most important and complete banks of the world data. The idea is to feed from these databases, the entries published by Wikipedia in various languages.

This new project aims to provide updated information to the entries that it is said “the free encyclopedia” as well as for anyone willing to use the data. They will be published on the Internet via Creative Commons license, one that, depending on the assigned permissions, allows reproduce, modify and edit the information provided for free.

The idea is very cool. For example, the entries on personalities could – all of them – to find data on date of birth and death of the individual date in wikidata, centralizing this information. A work less for those internet users who participate in the edition of Wikipedia.

The application goes beyond. Some entries in localized editions of virtual encyclopedia are not updated. With wikidata would, in theory, easier to keep these updated content with the latest information available, since a change in the database reflect the entries submitted in more than two hundred editions of Wikipedia available in the world.

The TechCrunch put his hands on the release schedule of wikidata. There will be three phases. At first, each Wikipedia entry wins an entry in wikidata to group links and basic information such as birth dates. The phase number two states that Internet users to add more details to the entries wikidata and use this information in Wikipedia – so the encyclopedia goes on to query the database for some data in the entry.

Perhaps the most promising phase is the third in which users will ask questions to the wikidata and the database engine (!) Will respond. For example, from the information would to say “which of the ten largest cities in the world have women mayors.” Get this result is difficult and involves the use of a lot of semantics to understand exactly what Internet users are looking for. Today Wikipedia to answer some questions of this kind, but are responses based on lists created manually. It is not an automated process.

The wikidata initiative has money from the Allen Institute for Artificial Intelligence, Microsoft co-founder Paul Allen. He gave 50% of the resources. It also has funds from the Gordon and Betty Moore Foundation and Google (get this), each bankrolling 25% of the 1.3 million foreseen in the budget. The engine certainly could make good use of a database of this magnitude – if wikidata go forward.

Who has to fear it is Wolfram Alpha, mechanism-based searches and calculations on mathematical operations. I do not know to what extent the two services are competing, but certainly wikidata can grab a slice of the requests currently made ​​in the search engine alternative .

The wikidata should be ready in March 2013.

I’m excited about what might be the wikidata. A database of this magnitude is lacking in our internets. And you, perked up too?