Freebase has been made public recently; it is a sort of Google Base with the goal to “[contribute to collect] data from all over the internet to build a massive, collaboratively-edited database of cross-linked data.”
Tim O’Reilly praised it [1] [2]; he had some thoughts about it and about the semantic web; he gave some opinions that leaded to a storm of blog post on planetrdf.com; Etc.
I will not enter in that debate. I have nothing to say about the Web1.0, Web2.0, Web3.0 or the NextWebVersion.NumberSomething except that these terms make people from around the World quite… unproductive. We have a system that let anybody write, publish and share documents on a space called the Web; and that since its beginning. So let see what we can do that such a system.
What I will do is telling you where I am with my vision of the Web, and to what it could evolve to. I will describe how the projects I am currently working on could make the Web different, hoping to make it better. I will only show you a schemas, with some explanations, of how I see the environment that such projects are currently creating; how users and developers will be able to use and contribute to these systems with the only goal of making the web open and better.
The next Web environment
Click to enlarge this schemas
1. Describing resourcesAs you imagine, the first actors of the system are Web users. The goal of these Web users is to describe things (resources). This behavior is not different from the past: Web users always described things; the only difference is that they have to use new methodologies, but even that, it is not always the case as we will see later.
Users will be able to describe things such as themselves, projects they are working on, relations between musical artists, albums, Etc. by using specialized software that will help them to describe these sort of things. Systems such as Wikipedia, Musicbrainz, Talkdigger, Livejournal, blog system using a SIOC Exportation add-on, Etc. (note: one could wonder why I name Wikipedia or Musicbrainz that doesn’t export anything in RDF; for them I would redirect them to the dbpedia and Music Ontology projects for more information).
But in the future, people will also develop specialized software that will help people to describe virtually everything.
2. Save descriptions
Systems will archive all these descriptions. Dedicated system for that task, some type of portals, personal web pages, specialized blog systems, specialized wiki systems such as the Semantic Media Wiki, Etc. These systems will publish the information to anyone who request it, exactly the same way as Web server publish web page content in HTML. The only difference is that it will use RDF instead of HTML to publish the same data.
3. Notify for new/updated descriptions
Most of these archiving systems will notify “notification [pinging] systems” such as PingtheSemanticWeb.com
That way, new/updated descriptions of something will be published to a multiple of web applications, software, crawlers, software agents, Etc.; requesting new/update descriptions from PingtheSemanticWeb.com
One of these systems is called Zitgist.
What is Zitgist?
In the past, I described Zitgist as a Semantic Web Search Engine. Great you will tell me, but what it means? I will refer you to this blog post I wrote a couple of weeks ago about what Zitgist is.
4. Send references of descriptions
PingtheSemanticWeb.com will send its pings to any system requesting the list of new/updated descriptions. On of these system is Zitgist. That way, it will be able to get the latest new and updated descriptions, and this, nearly in real time.
But any system can do the same.
5. Linked-Open-Data
This is a new project started by many Semantic Web enthusiasts, researchers and companies from around the World.
The goal of this project is to create a meta-database of interlinked databases such as Wikipedia, Musicbrainz, US Census data, DBLP database, Etc.
Such meta-database will be indexed into Zitgist to extend the descriptions its knows.
6. Other database
Other database, not part of the Linked-Open-Data project will also be indexed into Zitgist. In fact, any relational database can easily be converted into RDF and then indexed into Zitgist. One of the good examples of this is the conversation of the Musicbrainz.com database into RDF using the Music Ontology.
7. Describe resources
Zitgist is at the same level as any other application in this environment. So from its interface, Web users will eventually be able to describe things, and relations between these things, directly from its user interface.
8. Search
The more interesting feature of Zitgist is that it lets Web users searching in all that data. By using Zitgist, Web users are able to send queries such as:
- Give me the name of the albums published by Madonna between 1990 et 2000.
- Give me the name of the people that are interested in writing leaving near London.
- Give me the name of groups (group, organization, etc.) that has Brian Smith as member.
- Give me the name of the computer projects programmed using C++ that work for Linux or Windows.
- Give me the name of the discussion forums that are related to cooking.
- Give me the name of the cities in UK that have more than 150 000 people.
- Give me the name of the documents where its topic is a person named Paul.
- Etc.
9. Browse
Then from the results returned by Zitgist, Web users are able to browse the information about things. Things can be a:
- Person
- Project
- Geographical location
- Music artist
- Band
- Album
- Single track
- Etc.
- And relations between all these things
10. Possible duplications of Zitgist
None owns anything in that Web, so most of Zitgist could eventually be duplicated in other service. Why? For the only reason that most of the information it uses is available to anyone who wants to do something with it.
Examples
So, as you can see, this infrastructure enables a wide range of possibilities. In fact, a Web user could use Zitgist to find a thread in a Discussion Forum about the Semantic Web. Then he could browse the thread directly using Zitgist. If he has something to say on that thread, he could describe new facts about this particular thread directly on its blog. From there, its blogging system would publish the description to anyone who want it (like a normal HTML Webpage). After having published the data, its blogging system could ping Pingthesemanticweb.com to notify a wide range of applications that new descriptions has been published about a certain thread on a certain discussion forum. Then Zitgist would be notified of this new description and it will index it into its system.
After that other users will be able to see the new facts described by this user directly in the thread and Zitgist’s Web interface.
You could easily replace the subject of the above example by a musical artist, and change the blogging system for the Musicbrainz.com user interface.
Possibilities are endless… but the result is always the same: a distributed, open, meta-web. A sort of distributed meta-wiki created by a wide range of users, applications, services and systems.
Conclusion
This is my vision of my next web. I called it semantic web (as many other does), but you have the leisure to called it the way you want. I am developing Zitgist that should be released for a first round of users in the next few months; I am developing Pingthesemanticweb.com; I started the Musical Ontology; I am also participating to some projects like the SIOC ontology, the linked-open-data initiative; and this is the vision I have that lead my works.
I cited all the projects I am currently working on, but none of them could be possible without the tight collaboration of the semantic web community and the guys at OpenLink Software Inc. I owe them this vision and all my knowledge of the field. Thank you guys.
