Understanding Semantic Web (Part -4)

In this last post of the "Understanding Semantic Web" series (1 2 3)I will mainly talk about the current scenarios where Semantic Web techniques are being applied and what holds there in the future.

Semantic Web technologies are trying to penetrate every existing corner of software development, the way we build software from databases, middleware, applications to even designing UI for a portal. A number of databases (Kowari, rdfDB etc.), semantic middleware platforms (Intellidimension), multimedia systems are being built by the research community. In this post I will mainly concentrate on applications, which is where I believe the action lies!

In the previous posts I mentioned about ontologies which provide definition and are an important part of the application domain. Most of the applications that have become widespread are mainly with smaller ontologies that an individual can understand. The list below is not exhaustive and is provided to just give the readers (newbies) an initial impression.

User Applications

RSS - RSS is a widespread technology. Its commonly referred to as the "low hanging fruit" of the semantic web. RSS enables anyone to package content and metadata and provide updates. Mainly used for news and blogs updates.

FOAF - Friend of a Friend - FOAF provides vocabulary where friends can define semantic descriptions about each other. Simple GUI Applications exist for creating and visualizing FOAF data. People are trying to define a trust of web based on FOAF data. Perhaps this can revolutionize the way social networking works today or build smarter systems who know your friends and perhaps enemies too!!

Creative Commons - Creative Commons provides a very simple vocabulary to define licensing schemes for any kind of original/derived works (content, blog,website, pictures, audio , video ). An author can via simple form choose what kind of license he wants. A semantic description is generated at the back that an author can put on his website. Any search engine aggregates these descriptions and provides a semantic search. So I can now search for images which I as an author can use freely without paying a royalty. Yahoo recently incorporared creative commons in its search engine.

EventSherpa (SemaView) - Tool and service for creating and sharing events, schedules and calendar information over the Internet. (www.semaview.com or www.eventsherpa.com) I used their tool around a year back but apparently the site is down at the moment.


Enterprise Systems

Adobe XMP (Extensible Metadata Platform) - Adobe XMP Platform enables embedding of semantic descriptions as binary format within the multimedia (JPEG, GIF, TIFF images etc.) itself. So any image when once tagged via semantics using an Adobe Tool retains its semantics forever, for instance if image is copy pasted, the semantic descriptions get carried along. The platform is open to anyone to build or query semantic descriptions.


There are a number of other companies providing enterprise level solutions using Semantic Web Technologies e.g. Mondecca, Empolis, Ontoprise, Cerebra, Profium and others.

Application of semantic web technologies in the enterpise world is more in the sense of supporting a particular standard. The systems or platforms here offer a particular solution (Intranet or Internet portal which can be more dynamically configured or provide a better search using semantic metadata or as information integrators) for their clients. There isn't a standard software semantic web stack that is used (and is being built).


Research Applications

Haystack - information client as part of the Information Management Project at MIT. It aggregates RDF from multiple arbitrary locations and presents it to the user in a human-readable fashion.

TAP - A Project at stanford enabling Activity Based Search. Check out this.

Annotation Tools: Tools enabling rich annotations of HTML/multimedia documents. A number of such tools exist (MnM, Cultos,GATE, KIM, SWAN) enabling manual/ semi-automatic/ automatic generation of semantic descriptions.

AKTive Space Visualisations: Prototype showing geographically, research being conducted at different locations in UK. http://www.aktors.org/technologies/geography/


Research Themes: There are number of interesting projects going on in the research community.

- Languages - RDF, RDFS, OWL, RULEML (etc..)
- Semantic Databases - store and query RDF descriptions via SQL like query languages.
- Inference Engines - (inferencing RDF/OWL semantic descriptions and do reasoning/inferencing)
- Visualization Tools
- Ontology Core Research + Tools (Ontology creation, editing, merging, maintenance etc.)
- Buiding Domain Specific and Domain Independent Ontologies (Human effort for building ontologies by Ontologists)
- Semantic Search (Latent Semantic Indexing - how does one define PageRank in Semantic metadata ?)
- Semantic Middleware (standard J2EE/dot net, P2P , Asynchronous Messaging Services)
- Semantic Web Services (Service creation, discovery, query, integration)
- Semantic User Interfaces
- NLP (Natural Language Processing)
....
and the list is endless.


So what's the killer app of Semantic Web?

I think Semantic Web is much like the Web of today. Infrastructure is being built as of today but there isn't a single application like EMAIL which one could call a killer app. For an application to be killer, there has to be a widespread adoption. Semantic Web Techniques as of today are much heavy for a common man to comprehend and even for a software engineer to build. The revolution is more likely to occur in the middleware space which will act as an integration platform. For users, it could be an information dashboard, or a single service criss crossing multiple platforms -- the interface would likely remain the same or get simpler. What's going to change is the richness!

One important part within semantic web lies in the creation of semantics. With content creation moving from web onto mobiles, semantic web technologies can be clevery applied to bring in richer semantics. So if carefully harnessed next semantic web killer app might lie in the mobile world than on the web.

Semantic Web Technologies are not just about generating RDF/OWL encoded data, its also about being making systems more open. XML/Web Services have already started that trend. Companies like Google, Yahoo, Amazon, eBay (more recent ones like Flickr) provide a platform to query their data. And this has resulted in number of small startups and innovative services (Clubbing Craigslist classifieds with Google Maps).

In future one could envision every major service provider to start providing data and services as XML/Web services/RDF etc.

Monster providing Jobs data.
Social networks like LinkedIn, Orkut.
Match.com
Travel Sites

As each of the above services start opening up, one would see a greater ease in information integration and newer services propping up. A killer application might be the one which helps one manage himself/herself better. RSS removed the pain that one doesn't need to go and visit every website even though one is essentially viewing the same content. The semantic web would remove the pain more or less in the same way as RSS did with content. One thing it might fuel more is more innovative services and more Innovation!