Collaboration

Cheat Sheet: The semantic web

The internet is evolving...

What's this all about then?
The semantic web is being touted as the next stage in the evolution of the internet, no less.

I thought that was web 2.0?
Web 2.0 was all about the new generation of online applications - like Facebook and YouTube - whereas semantic technology is about making online information and services more useful.

How so?
Computers at the moment merely read the HTML documents that make up the web, they don't understand them. For example, if you put a term into a search engine, it returns the results that match the keywords you put in. The semantic web however helps the computer to understand the purpose of your search - the context or meaning around it, not simply the individual words themselves - and returns results accordingly.

The semantic web works by adding the context of data through the use of tagging - or metadata - in order to make information on the web understandable to computers.

As more data is tagged in this way, search will become more accurate and useful and the information people are looking for is more likely to appear. The semantic web means search engines won't simply be relying on keywords or explicitly stated information, they'll be able to mine metadata to distinguish between types of information - events, music, e-books, relationships between people - to get better results.

Formal specifications have already been developed for semantic tagging, including Resource Description Framework (or RDF) and Web Ontology Language, both of which provide descriptions of terms and concepts that can be applied to data in order to add context.

Tech Hotspots: The list

1. Silicon Valley
2. Bangalore
3. London
4. Tokyo
5. Boston
6. Cambridge
7. Shanghai
8. Tel Aviv
9. Seoul
10.Beijing
11.Chennai
12.Pune
13.Singapore
14.Helsinki
15.Moscow
16.Hong Kong
17.Hyderabad
18.New York
19.Sydney
20.Shenzhen

How do you go about tagging data then?
An example of semantic tagging could be an online sports clothing company wanting to more accurately catalogue its products. Each product could have several RDF tags such as the colour, style and associated sport. These tags would provide context to other information within the catalogue, which could be for example, red, boot, football.

With all those metatags, someone looking for the football boot of their dreams is more likely to find it when they trawl the company's website.

So semantic technology helps improve search but what else can it do?
It's not just search engines that can use this type of information - the online retailer could use the tags to help structure its databases and integrate that data across a wide range of applications more effectively.

In the longer term this kind of technology could be applied to improve things like service oriented architecture and make it easier to search for online services in the cloud.

With SOA for example, when looking to link up software within a business the tech team can link separate applications using semantic tagging, instantly associating them with each other within the architecture.

So if an HR database needs to be connected to the central accounts systems, it could be tagged appropriately to direct people in the right direction when required.

Improved search for cloud services would mean individuals and businesses could locate and acquire applications that most accurately match their requirements much more quickly.

Who says it's such a big deal?
Well Tim Berners-Lee - widely regarded as the founding father of the world wide web - is a big fan and has been working hard to spread the word. And you'd assume he must know what he's talking about.

He's not alone as other academics recently joined forces with Berners-Lee to set up the Web Science Research Initiative (WSRI) aimed at promoting semantic web technology into reality.

Indeed at the launch of the WSRI, Google CEO, Eric Schmidt, said his company would be interested in employing people trained in semantic technology.

It still sounds a bit pie in the sky to me…
Semantic technology is certainly something that is fairly niche at the moment, being taken up by fairly specialised user communities, but the signs are it won't stay that way for long.

Recently, semantic web technology has started to emerge from the R&D labs and is being used for some real and useful work. There are some big names looking at the possibilities of semantic web too.

Like who?
Newswire Thomson Reuters recently launched its OpenCalais service that allows other companies to make use of semantic technology.

Originally developed to boost the company's algorithmic trading capabilities, the service can be used to add semantic functionality to online content such as blogs, content management systems, websites or applications.

Another big name to use semantic tech is Yahoo! that has developed its SearchMonkey service. Like OpenCalais, this application can be used to make web content more useful and visually appealing by adding semantic information.

Is that all?
Nope - Microsoft is investing in semantic web technology after acquiring natural language search firm Powerset for $100m in August this year.

BT is also developing and using its own technology aimed at helping the telco more effectively interrogate its various databases through a simple search query. BT is already using this tech in its sales and billing departments and plans to roll it out to its business customers in the near future.

There are certainly some big players there but is it really going to catch on?
That's the million-dollar question and there are some people that doubt whether the semantic web can really be a genuine possibility.

Canadian blogger Cory Doctorow, the man behind BoingBoing, coined the term "metacrap" - metadata that can be misleading. Doctorow says this could be for a number of reasons including the fact that people are lazy or there's more than one way to describe something.

Others say the limitations of software engineering is an obstacle as it means the extra layers of data that are required may not in reality add as much value as it has been suggested.

So the jury's still out?
You could say that but if Tim Berners-Lee is touting it, it would take a brave person to bet against semantic web taking off...

Editor's Picks

Free Newsletters, In your Inbox