Faviki: social bookmarking for 2010
By bblfish on Jan 13, 2010
Faviki is simply put the next generation social bookmarking service. "A bookmarking service? You must be kidding?!" I can hear you say in worried exasperation. "How can one innovate in that space?" Not only is it possible to innovate here, let me explain why I moved all my bookmarks from delicious over to faviki.
Like delicious, digg, twitter and others... Faviki uses crowd sourcing to allow one to share interesting web pages one has found, stay up to date on a specific topic of interest, and keep one's bookmarks synchronized across computers. So there is nothing new at that level. If you know del.icio.us, you won't be disoriented.
What is new is that instead of this being one crowd sourced application, it is in fact two. It builds on wikipedia to help you tag your content intelligently with concepts taken from dbpedia. Instead of tagging with strings the meaning of which you only understand at that time, you can have tags that make sense, backed by a real evolving encyclopedia. Sounds simple? Don't be deceived: there is a huge potential in this.
Let us start with the basics: What is tagging for? It is here to help us find information again, to categorize our resources into groups so that we can find them again in the rapidly increasing information space. I now have close to ten years of bookmarks saved away. As a result I can no longer remember what strings I used previously to tag certain categories of resources. Was it "hadopi", "paranoia", "social web", "socialweb", "web", "security", "politics", "zensursula", "bigbrother", "1984", ... If I tag a document about a city should I tag it "Munich", "München", "capital", "Bavaria", "Germany", "town", "agglomeration", "urbanism", "living", ...? As time passed I found it necessary to add more and more tags to my bookmarks, hoping that I would be able to find a resource again in the future by accidentally choosing one of those tags. But clearly that is not the solution. Any of those tags could furthermore be used very differently by other people on delicious. Crowd sourcing only partially works, because there is no clear understanding on what is meant by a tag, and there is no space to discuss that. Is "bank" the bank of a river, or the bank you put money in? Wikipedia has a disambiguation page for this, which took some time to put together. No such mechanism exists on delicious.
Faviki neatly solves this problem by using the work done by another crowd sourced application, and allowing you to tag your entries with concepts taken from there. Before you tag a page, Faviki finds some possible dbpedia concepts that could fit the content of the page to tag. When you then choose the tags, the definition from wikipedia is made visible so that you can choose which meaning of the tag you want to use. Finally when you tag, you don't tag with a string, but with a URI: the DBPedia URI for that concept. Now you can always go back and check the detailed meaning of your tags.
But that is just the beginning of the neatness of this system. Imagine you tag a page with http://dbpedia.org/resource/Munich (the user does not see this URL of course!). Then by using the growing linked data cloud Faviki or other services will be able to start doing some very interesting inferencing on this data. So since the above resource is known to be a town, a capital, to be in Germany which is in Europe, to have more than half a million inhabitants, to be along a certain river, that contains certain museums, to have different names in a number of other languages, to be related in certain ways to certain famous people (such as the current Pope)... it will be possible to improve the service to allow you to search for things in a much more generic way: you could search by asking Faviki for resources that were tagged with some European Town and the concept Art. If you are searching for "München" Faviki will be able to enlarge the search to Munich, since they will be known to be tags for the same city...
I will leave it as an exercise to the reader to think about other interesting ways to use this structured information to make finding resources easier. Here is an image of the state of the linked data cloud 6 months ago to stimulate your thinking :-).
But think about it the other way now. Not only are you helping your future self find information bookmarked semantically - let's use the term now - you are also making that information clearly available to wikipedia editors in the future. Consider for example the article "Lateralization of Brain Function" on wikipedia. The Faviki page on that subject is going to be a really interesting place to look to find good articles on the subject appearing on the web. So with Faviki you don't have to work directly on wikipedia to participate. You just need to tag your resources carefully!
Finally I am particularly pleased by Faviki, because it is exactly the service I described on this blog 3 years ago in my post Search, Tagging and Wikis, at the time when the folksonomy meme was in full swing, threatening according to it's fiercest proponents to put the semantic web enterprise into the dustbin of history.
Try out Faviki, and see who makes more sense.
Some further links:
- Auto-Tag Delicious Bookmarks and Share Them Easily On Twitter With Faviki a blog that goes into more detail on Faviki.
- Faviki's Feed of resources tagged with Faviki.