Internet searches have the potential to be infinitely more effective, with a little innovation and a lot of collaboration, as various Irish initiatives have discovered, writes CLAIRE O'CONNELL

HAVE YOU ever gone looking for something online and ended up facing acres of rubbish, through which you then have to painstakingly pan in search of the gold?

Or maybe you’ve felt your frustration levels well up because, no matter how many search terms you fling out into the darkness of the web, you still can’t hit the target. You’re not alone.

But one person’s adversity is another’s opportunity: the explosion of information across the internet has left the door open for new approaches to help you find what you are looking for, store it in a way that’s of some use and even socialise online while you are at it.

One company that hopes to improve the signal-to-noise ratio for users is HeyStaks, a spinout from University College Dublin. Its chief scientific officer, Professor Barry Smyth reckons their approach – which allows you to organise, store and share your web search results in folders or “staks” – could substantially cut the time wasted while searching online.

“A [web] user such as a student might be spending 15 to 20 hours a month just searching for things, and sometimes up to half of that time is wasted,” says Smyth, who holds the digital chair of computer science at UCD. “We can recover a lot of that.”

HeyStaks is based on research carried out at Clarity, a Science Foundation Ireland-funded joint initiative between UCD, Dublin City University and Tyndall National Institute in Cork. Ongoing studies there show promise that their “stak”-sharing approach makes web searches more effective, says Smyth, who is also Clarity’s director. “Because people can collaborate with other people as they search, they are finding things they wouldn’t normally find,” he says.

HeyStaks is currently in closed beta as early users help refine the approach, but over coming months we should all be able to download the toolbar, start building and sharing our own search staks and eventually synchronise staks across multiple devices.

A spinout from the University of Limerick is also trying to make online searches more productive – this time for a more tailored audience of prospective students who are trying to navigate the maze of available educational courses.

“Not too long ago, if you wanted to get professional qualifications, you had to go to a college or professional training provider, but in the last few years the sources of where you can get knowledge has exploded,” says Dr Brendan Cleary, technology officer with LearnOpt.

“So how do you help people navigate through that quite confusing place, if you are not in it all the time? We were looking for ways to bring the training that would be relevant to the learner, instead of having them go out and search for it.”

The resulting technology, which underpins LearnOpt, grew from a Higher Education Authority-funded project at UL’s Enterprise Research Centre that was geared towards helping direct people to suitable courses.

LearnOpt uses a general recommendation approach that might be familiar to users of large online retail sites, explains Cleary. “Those systems mine the intelligence of the users on the site – the more you look at the more it builds a profile about you. So we thought, ‘could we apply that to learning?’ – [because] a course or degree programme, they’re just pieces of information,” he says. The user can then build their own, customised portfolio with the information they find.

LearnOpt is already up and running at the site bluebrick.ie, where users can search and apply for part-time and evening courses on offer across institutes of technology in Ireland, and the company is now approaching other learning institutions both at home and abroad, according to Cleary.

Another niche market for an improved online service is the scientist who wants to order a product, as Dr David Kavanagh found out first hand when he was doing his doctorate in biochemistry at UCD.

“Scientists have to look through catalogues or look at 20 different websites, not really knowing what the best products are for the market,” he says. “It’s a cumbersome process and big organisations have no infrastructure on how to gather information about the performance of purchases.”

Meanwhile, suppliers can find it difficult to get access to their customers, adds Kavanagh. “They put together fancy brochures, put things on notice boards, try to make appointments with them, but it’s a very crude process,” he says.

So a new service, Scrazzl, is being developed to link everything together, creating a network for scientists and suppliers that allows people to shop for products, review them and share information.

“A scientist can find what they want on the system, like Amazon or eBay, and because you create a network within the system there’s a real capacity for sharing information,” says Kavanagh, who is managing director of the UCD spinout.

The idea is that scientists can not only review products and find out what others think, they can also forge links with people working with the same materials, explains Kavanagh. And suppliers can not only build relationships with their customers without getting on the road, they can also see which products are performing well.

“We have built a system that can be dropped into any organisation or department across multiple institutes and it allows scientists to start collaborating,” he says. “And it speeds up innovation on the commercial side of science, because the companies will get feedback quicker and more accurately on the products.”

Suppliers would pay a fee to be part of the Scrazzl network and access statistics, which is an obvious source of revenue, explains Kavanagh.

But for free search tools such as HeyStaks, the question of the bottom line is a little trickier, says Smyth, who suggests that the first approach is to build up user loyalty and understanding and to capture their attention.

“It sounds strange, but it’s not all about running after the revenue on day one. We need to understand what’s best for our users, what they want and then figure out how we can wrap some revenue-generating services around that.”

Companies such as Twitter have done this well, notes Smyth, by keeping the core services free for the benefit of their users and bringing in revenue through secondary services.

“Search is a huge market,” he says. “It’s the one internet market that really has proven to be a major revenue generator, and any new search technology that can disrupt that market has the potential to be very valuable – even though it’s not always clear exactly where the revenue comes from.”

Gettign semantic with the web

THE "SEMANTIC web" is an emerging term you may have heard of and smiled politely while you wondered what it was all about.

So, what is it all about?

"The main idea is to make the web smarter," says Dr Alexandre Passant, a unit leader at the Digital Enterprise Research Institute in NUI Galway.

At the moment, the web is just linked documents and, while humans can spot connections that link datasets, computers aren't that way inclined, explains Dr Passant.

By describing objects in a page and linking them, you can make online datasets more connected and useful: "You can just add a few markers in your page – this one is about a person, this one is about a city, this one is about a university – it depends on the granularity and level of the information you want to give," he says.

Using semantic web and linked data technologies, Passant built a music recommendation system that recommends bands to a user based on features of a band that they already like, for which he won an award at the recent Extended Semantic Web Conference.

The semantic web is starting to grow and already governments and companies are starting to link datasets in this way, according to Passant. "There will be more and more websites [with this]," he says.