Find the Wikipedia Map Interface here: http://km.aifb.kit.edu/sites/wikipediamap/

Heise online and taz.de reported about Wikimedia’s plan to build a centralized database that will be shared over all language versions of Wikipedia where the RENDER project is involved.

Read full article on heise online and taz.de

Eli Pariser and other critics think this is dangerous and believe that this approach prevents us from seeing and using the full potential of the Internet, not being presented with information that doesn’t fit into our own universe of opinions and interests. Eli Pariser calls this “invisible autopropaganda, indoctrinating us with our own ideas”. In his book “The filter bubble: what the Internet is hiding from you” he goes into detail how such a filtered Internet can be dangerous and how sites can give users more control over their personal data.

A click of the button will result in the launch of a Love Interface, in which the user is presented a number of options for what kind of love to send and images to append to a free-text compliment area. The resulting declaration of support is published to the receiving user’s account discussion page and is monitored by top Wikipedia users to prevent misuse.

With this release, Ontotext have rebranded their line of OWLIM editions and introduced a new member of the family:

• OWLIM Lite is the new name of SwiftOWLIM
• OWLIM SE (Standard Edition) is the new name of BigOWLIM; it is positioned as a single-server product,
• OWLIM Enterprise is an extension of OWLIM SE, which includes the Replication Cluster capabilities. It is priced and sold as a separate product intended for scalable, high-resilience, mission-critical installations, offering extended online re-configuration and other features which improve manageability

Scalability to 1 trillion nodes: A brand new indexing mechanism has been developed for OWLIM SE that uses a configurable bit-size for internal identifiers of nodes in the RDF graph (URIs, blank nodes and literals). For very large datasets, the bit-size can be set to 40 bits, allowing up to 1 trillion unique entities to be stored.

Faster SPARQL 1.1 Query support: Thanks to Ontotext’s investment in development resource for the Sesame project, all OWLIM editions now boast support for this evolving SPARQL standard. While handling of SPARQL 1.1 through Jena was fast enough to allow OWLIM to get excellent scores in the latest BSBM evaluation, using SPARQL 1.1 through Sesame makes it even faster and available in all editions of OWLIM. SPARQL 1.1 Query specification offers new features such as: Aggregates, Subqueries, Negation, Expressions in the SELECT clause, Property Paths, Assignment, Short form for CONSTRUCT and an expanded set of functions and operators. Federation will be supported in a subsequent release.

Easy server deployment: every edition of OWLIM includes re-packaged Sesame Web applications for easy deployment of OWLIM servers. Now OWLIM can be deployed simply by copying a WAR file. New OWLIM repository instances can be created using the Sesame Workbench UI administration tool.

There are number of other new features for specific editions:

• OWLIM SE now provides a remote notification mechanism that includes transaction information.
• Performance analytics: OWLIM SE provides information on cache utilization and behaviour through a JMX interface. This feature can be useful when investigating loading behaviour and query performance.
• OWLIM Lite is now packaged as a single jar file, with a simpler free-for-use licensing scheme.

Note: This beta release can not be used with existing OWLIM storage files. The full release due in couple of weeks time will include an automatic upgrade facility for updating files created using previous versions.

More information on OWLIM can be found here: http://www.ontotext.com/owlim

The algorithm determines from previously visits to profiles and pages, the number of comments a user creates, user’s visits to profiles/pages etc. how relevant a post is. Simply said, people that like cats will no longer see photos of dogs but are rather presented with pictures and posts from other cat-lovers. Doesn’t sound that bad, but it might become a problem when people use social platforms more and more as their primary sources to get news and opinions. Irrelevant or contrary opinions are simply not displayed and that can create a wrong picture of the world.

Eli Pariser, former Executive Director of MoveOn.org, and the organization’s current Board President, talks about a world where we only see things on the Web of which algorithms have decided that we must see these things – but we will not see what we should see. The article’s author raises another problem: It may seem plausible that people in a specific environment hold back with their opinions. If that happens we lack a different viewpoint – and the digital world seems much more unanimous than it truly is.

Lesen Sie den ganzen Artikel hier: URL| Download PDF (No hits, 92.73 kB)