We have IMDB, we have Last.fm, we have Wikipedia and thousands of other database websites with information about media items. Currently we search for information on books, music and video via search engines like Google, Yahoo or Bing. Searching through website that hold all the information is done very well. But the raw results are not always satisfying enough.
Are we looking for a website like last.fm when we are looking to buy an album? I know Google’s shopping filter would be an option for that. But what about lyrics, videos, publishers, local shops and all other kind of data one could be interested in. Last.fm provides a lot of data but with an agenda. Not always ambiguous, focused on music alone and not very singular.
Libraries do have this kind of structure and ambiguity. What they are missing most of the time is completeness of data and the simple fact that it is only data from one library. Can we create a more universal database focused on providing single records of data that contain everything one would ever want to know about it?
The library model
I’ve been working with some library data lately and I found that they have very interesting ways of structuring data. Larger libraries yield a set of records to show up for different locations. They link different versions of a book to show up as one record. This is a very structural and orderly way to show the data. Does a harry potter book need to show up more then once in a result list of would you just be content with a list with all the books in the series once?
The library model might be local and based on a single instance. But what if we take the single record model with a local touch we might have something to work with.
Single record model
Using a single record to collect information on an item is basically adding a layer to search results. Instead of finding 545.000.000 results when looking for harry potter you will only get 7. Every book in the series is it’s own record and the movies can fall under the same flag. If we look at the content of the book versus the content of the book we will see that they are representing the same thing. It becomes a bit more tricky when looking at sound-tracks for the movie but in general we can connect all these items to one single source.
Another advantage of the single record model is that you have one area to show meta information like different versions, languages and editions. Information like origin, creators and publishers can then be easily accessed. Instead of having to plow through all those different results it is all captured into one nicely formatted record.
Local over global
In the modern day and age if we talk about information in media we will almost always talk about global information. IMDB does not only hold information from internationally published movies and tv series. Also the local television and cinematic content can be found in this database. Maybe not as complete as the international the local touch gives it that extra completeness.
Local and global both do have a very different impact om people. global news can be seen as just an event. Many are watching the developments in Libia from a distance. It is not a very tangible thing but we do want to give support to the rebels. Local new has a more personal impact. A local fireman being killed has much more impact than one being killed a few towns over. The more global you go the bigger the impact must be before you can get a feel for it locally.
The library has a very local impact and therefore takes local content and promotes it over all the more globalized content. As they are providing this service they keep an eye out at the balance between global and local information. This is very useful in a big media library. Local content should be presented before the more internationally oriented information. In the Netherlands it is less interesting that the Millennium books have a French version of the book. It is a lot more logical that in the Netherlands you get presented with the Dutch and Swedish version first as the are the local and the original language of the book.
Value of information
The library is something more than just a place that holds books. It has a certain value as it safeguards the value of information. So, what is that hidden value of information? Why do we not accept wikipedia as “real” information and we do accept a paper encyclopedia?
It is because of the filtering and centralizing information. It is because we trust someone to safeguard the information value for us. We trust them to give us unbiased information with some recommendations. We value the library as it safeguards the value of information.
Using the powers of the internet
It has been said in the past: “This internet thing is going nowhere” or “The internet is just a passing fad”. Both came to the forefront about 15 years ago. Now 15 years later we know better. The internet has revolutionized the way we the way we experience media, the way we work and even the way we live. Information freedom is something that is becoming part of the law and some internet companies are part of 100 most successful companies in the world. Even the revolutions in Egypt and Libia are accredited to the rise of social media.
So what are the real “powers” of that internet thing and how can a glocal universal media library benefit from it?
Look at wikipedia. Thousands of people are adding and updating information in the internet. One minute after Steve Jobs stepped down as Apple Inc. CEO it was already added to his wikipedia page and the new CEO was placed on the Apple Inc. page. Community can work for you to update and add information that you might be missing.
There are also problems with this way of working. Information might be wrong and some information might be based on opinion or does not give a full image. Mechanisms to keep that information into check are required to deal with these kinds of situations. This is also the reason why we trust libraries more than just the internet. It is a web of trust we create with these local “experts”.
To use the best of both worlds we need to take those local experts and let them either verify or gather information. Using their knowledge and expertise they can determine what is what. Community of random people is also important. But starting with a community of experts and adding “unverified” information from other can be a good community model.
I’ve already mentioned places like IMDB, last.fm and wikipedia. But there are many other sources that are open and ready to use. This way we tap into the content released by publishers, community content and all other sorts of information. Services like LibraryThing have a vast database with reviews and ratings for books. Adding that information from those services can really enrich the experience.
Local vs. global
We talked about the value of local information. It’s interesting how locality works. In many ways locality is not about actual location. But about finding a connection between people. If i’m on a subway in Amsterdam and I meet someone from my hometown I greet them even though I would barely know them. If I were in NYC on the rube I would greet anyone from the Netherlands… Just because we have a connection. This is an example about location… but how about other things? I have a t-shirt that says: “Skynet development team” and when I go to the cinema I have an instant connection to people who like the terminator movies. Just because we both have the same interest or do the same thing. In a globalized world we look for a connection to our own “tribes”.
In the world of social research “tribes” are a common element. Wearing the same brand shoes, listening to the same music, playing the same games or even the way you walk the street determines what tribe you belong to. But tribes are again nothing more than just a connection between people. A random person will belong to several tribes at the same time, and even more during the course of once life. But how can we use the power of tribes?
The phenomenon of tribes are already being used in the world of search and social media. We get recommendations on our previous actions. These actions match together in a “tribe” and this allows for recommendations based upon your personal preferences. If I search for “sexy python” I will get a lot of links to the sexy framework in Python. I will not get Britney with a python around her neck… My tribes are probably tech, design and gaming.
Location does matter
When looking at media locality actually really matters. The over-complicated rights in media alone deserve filtering on locally available media. But even in other situations. If in Sweden I will probably would like to read the books of the Millenium series in Swedish. In the netherlands I would probably would like to read the book in either Dutch or maybe English. Swedish is less relevant outside of Sweden so location is key here.
Locality can also be a physical thing. I’m looking to buy a Blu Ray disk of the latest harry potter movie. I can order it online but I need it right now… Local stores might have it but I would need to visite a couple of them to be sure they have it in stock. What if I search for the movie and find local stores that have it in stock? Interesting?
Proposal: What if…
What if we create a central database with all media content. This can be indexed, scraped or whatever, but let’s say we can do this. Connect all the content of one specific item like the matrix reloaded, it’s movie score and it’s soundtrack should be connected. We connect items in a series like the Millenium books and movies.
So now we have this vast database of connected content. Create a search engine to start using it. Search on the items, and the content that comes with it. Use external data to enrich it. Think locally as well as globally while enriching the record for a specific user. Use the expertise of the user but also draw in specialists to maintain all the extra information about an item.
What you get is a very complete searchable catalog of media items. Using the power of the library and the power of the internet to allow users to get a personal and glocalized experience.