A question of semantics


24 Jun 2004

The Digital Enterprise Research Institute (DERI) is a collaborative venture between National University of Ireland Galway, the Leopold-Franzens Universität Innsbruck and Hewlett-Packard (HP). It was set up last year with funding from Science Foundation Ireland (SFI). Last month, DERI announced a e12m research programme, again funded by SFI, to develop next-generation web technology. Shortly after the announcement, silicon republic.com spoke to Prof Christoph Bussler, vice-director and semantic web services cluster leader at DERI, Galway.

Siliconrepublic.com: There is a lot of talk about the ‘semantic web’. What exactly do you mean by that term?
Christoph Bussler: The example I use a lot is, if you have never been in Galway but you want to know about it before you come, then you would go to a search engine and you would type in ‘Galway’ and you get back every webpage that contains the word ‘Galway’. This might or might not be relevant. But you can’t tell the search engine “Only show me what’s relevant”.

The only thing you can do is type it in and hope that what you want comes back. You then have to look at every single result and make a judgement as to whether it is relevant or not. So then you find out there are two cities called Galway, one in Ireland and one in upstate New York. But there is no real way to tell the system I only want the Irish one. But now let’s assume the semantic web is in place. Then you couldn’t just type in ‘Galway’. You would have to say “I’m looking for Galway, which is a city” and then, depending on how familiar you are with it, you would say “I’m looking for Galway in Ireland” and the search engine would look just for those things, because it would know that there are things called ‘cities’ and that cities are in countries, so it would have the semantic information that it would use to accept your instructions. From an end-user perspective the only additional thing the user has to do is say “Galway is a city,” which the end-user knows anyway because he is looking for the information. But underneath, the technology to make this happen is a lot of work. But the beauty of it is, for the end-user, it does not change a whole lot what he has to do, but he gets the benefit in terms of the precision of the data coming back. If there is only one answer to a question you only want one answer back and not 5,000, which is what everyone experiences every day. This is one pillar of our work.

SR: What is the other pillar?
CB: The other pillar is called semantic web services and this is a technology that allows for the dynamic electronic interaction between businesses. So that the communication of things such as purchase orders, shipment notes, invoices and so on that normally happens on paper can be done electronically. You can do this already but if you want to set up two businesses to talk to one another it requires a huge effort and can take up to a year. Mostly what these programs are dealing with is the semantic mismatch, the imprecision of what terms mean. For example, if you ask me for a quote and I send back a price, then you might think this price is inclusive of Vat, but I might think it is exclusive of Vat. If you don’t clear up this misunderstanding then you are going to have a big problem down the road. So if the systems say Vat is included and you understand Vat is included then there won’t be any dispute.

SR: But isn’t this already catered for in things such as universal business language?
CB: No. All of the current efforts give you languages to write things down but they don’t give you a way to precisely define the semantics.

SR: How do you plan to commercialise the research?
CB: We have a business development person, whose sole job is to build up a relationship with every single company in Ireland interested in working in this space. He is an entrepreneur himself so he knows how things work from the companies’ perspective. Step one is to build up a network of companies so you know what’s out there and what areas they are interested in. More specifically, we are meeting with companies to explore forms of collaboration. This might involve a company placing an employee with the institute for a year so they learn what we are doing and at the same time he talks to researchers about what he knows so it’s a two-way learning situation. Other companies might simply want to have a consulting activity while other companies are just curious and are looking for new possibilities. There is nothing we can’t talk about.

SR: What sort of timescale are we talking about? Do you think we might see the beginning of commercialisation of your research in a year’s time or two year’s time?
CB: Yes. The institute was started less than a year ago but the people we are hiring are in this space a lot longer. So they bring with them ideas that we can bring to a proper level of maturity right away and as new people come on board they bring with them new ideas so it is a rolling effect from the beginning. Our scientific director has been in this space forever. I come from the integration side and I have been in this space for 15 years now. And HP in Galway, our industrial partner, has been in this space for a long time, although in this case more from an application side. Nevertheless, because it has done it HP knows a lot about what we have to achieve.

SR: I understand you have a schools outreach programme — DERI-Óg. Can you tell me a little bit more about this?
CB: Right now the programme is aimed second-level students, but we may go to primary level with it. Here you wouldn’t go explaining what the semantic web is but you would kind of make sure that the children understand the web. It is a sort of pre-development stage and then at second level go more into the real stuff. We are trying to develop a tool on the web so that children can work with the tool and explicitly build a semantic web from their viewpoint.

But the more important element from Ireland’s point of view is that children must be interested in research. A lot of kids think research is an old guy with a beard and a white coat. In fact it is exciting. You want to show them that research is something they might be interested in. So it is very important to make them excited about research and science in order for them to be the next generation of students. There is a whole ecosystem here that we have to put in place from the intake of children all the way to business.

By David Stewart

Pictured at the launch of the Digital Enterprise Research Instititue is Tánaiste Mary Harney TD