Semantic Web Company

With David Huynh’s Freebase Parallax, an inspiring new user interface has come out, which conveys a sense of a future where googling isn’t the ultimate way to find information on the web. David was also strongly involved in some projects of CSAIL at MIT which also dealt with the “simple” question of making the (web of) data more accessible for users who aren’t aware of SPARQL, SQL or OLAP cubes. For instance, Exhibit has become a widely adopted environment for faceted search within a given dataset. A bit more sophisticated is the nested faceted browser – and now there is Parallax, too.

In his screencast about this novel browsing interface, David stresses the advantages over Google or Wikipedia. Is this a fair competition? What is it about Parallax that makes it a choice interface and when?

1. If you want to learn a little more about Abraham Lincoln – go to Wikipedia!
2. If you want to know where you can find even more information about Abraham Lincoln – go to Google!
3. If you know a bit about Abraham Lincoln already, and you want to aggregate or compare some facts of his life to that of other presidents or if you want to visualise data on a time-line or on a map – learn how to use a tool like Parallax (and don’t complain if it isn’t as simple as Google anymore)!

In some respect, Parallax is the building block that has been missing in the web universe: (Professional) fact finding on the web could work like this in the close future (although Parallax needs some more servers and – indeed – more data in its database).

But what if Parallax became the graph-based UI on top of Freebase + LinkingOpenData?

Refined search results for “Shaka” on Parallax

After publishing the first Cuil post this morning and adding it to my list of semantic search engines, I realized that I hadn’t checked its foreign language abilities. German is the only language I can assess as a native, so I searched for a few popular German terms like Fernsehprogramm (TV guide), Thomas Godoj (German winner of Pop idol) and Bauchtanz (belly dance). Turns out foreign languages might (yet) be its Achilles heel…

First off: Cuil (not quite understandably!) does not know Thomas Godoj, it doesn’t even recommend it as a popular search. Not even his official homepage is listed. Cuil promises to respect privacy – which might mean that it doesn’t even track IPs because the single quality search result it offers is the Thomas Godoj entry in the ENGLISH wikipedia. The selected pictures show pop singer James Blunt and the World Cup Trophy (from 2006 – i.e. before Thomas’ reign) and most search results actually link to video pages where Cuil probably found the most relevant tags as the links read:
video.worldcupblog.org/tags/Godoj/
www.mefeedia.com/tags/godoj/
www.bitdig.com/search/torrent-thomas+godoj/

On to Fernsehprogramm… (more…)

Not a fortnight seems to go by these days without the announcement of another new Semantic Search engine – hence I though I sit down and draw up my own little list of currently available search engines. The amount of semantics in them isn’t always transparent – hardly any of these search engine providers wants to disclose the ingredients in their recipe. I’ve also included a few search engines or search engine type applications that rely on collective or social intelligence to improve their search results.

If you have heard of any other semantic search engines that are not yet on the list, please leave a comment. They appear in alphabetical order, i.e. in no particular order. The information contained in ‘Notes’ is not intended as an independent evaluation. You might also want to check out the Top 100 list of alternative search engines on ReadWriteWeb – even though a number of search engines – sadly, are no longer online since the article was published in January 2007… (more…)