The need for semantics

What is the role of semantics?

I think we need to disambiguate "semantic technologies" and the "semantic web." The semantic web is a vision, championed by Tim Berners-Lee and others, whereby websites and databases are able to better exchange data and interoperate by using RDF triples. This is a great vision but has little to do with working with semantic technologies to build information products today.

Semantics refers to technologies and processes that assign meaning to content in ways that computers understand. The purpose of using these technologies is to support features and functionality in online information products and to help build them more efficiently.

There are a lot of theory and buzzwords being used with regard to semantics but very little in the way of features and functionality that people find useful. At Silverchair we use semantic technology – not because it is a hot technology but because it is the best technology to support real-world cases in our applications.

We started using semantic technologies to solve the problem of the book index. In the late 1990s we started developing online textbooks and realised we needed a repeatable and scalable way to index them.

This led us to our first work with structured vocabularies, and eventually to the development of the polyhierarchical taxonomies and tools we use today to develop dynamic topic collections, help learners focus their studies, provide a more effective search experience and provide a more personalised user experience.

Although the industry is still learning how to apply semantically-enriched content in applications, it will become so fundamental that we stop talking about it, in the same way that we don’t really talk much about XML anymore, because it is just assumed to be there in the background.

Ultimately, semantic enrichment will allow people to make better use of their time. People have too much information. Semantic enrichment presents data in a way that computers can interpret.

Another big buzzword is workflow integration. Students have a workflow involving textbooks. Researchers have a workflow that might include looking at laboratory documentation. Clinicians have a different workflow involving information about patients’ conditions. For different groups, the ways of keeping up with information are very different. People need the right information presented at just the right time.

Semantic technology can help and we are starting to see more and more of it in workflow integration. For example, the new ClinicalKey from Elsevier makes a lot of use of semantics. We do likewise with 5 Minute Consult, which we developed for Wolters Kluwer.

How will journals change?

People have been predicting the end of journals for a many years but I don’t see that on the horizon. We’re still publishing articles in journals and the article is, and will remain, a pretty good structure for the kind of information they report. Moreover, the journal serves many different – and essential – functions and information dissemination is only of them. Journals are important filters – both of articles and for career advancement. Most of the experiments aimed at changing this end up reinforcing these filtering functions.

What does change is the ability to interrogate articles better and provide more useful information and services to the reader. Journals will become better filters and will interoperate in a more complex information ecosystem that is still emergent. When the internet and web first came along they emerged from research entities, so were fairly comfortable for STM publishers. More recent waves of technological innovation, however, have come from the consumer space.

STM publishers have had a harder time adapting to these consumer-born technologies. In some cases, like with mobile technology and clinical medicine, publishers have been racing to catch up with what turned out to be an incredibly rapid adoption by physicians.

In the case of social tools, publishers have been experimenting without much in the way of result. Many social tools were created to do things that STM organisations had long ago figured out.

Journals already include commenting in the form of letters and peer review, and researchers already build communities when they meet at conferences. A lot of time and money has been spent trying to retrofit social tools to professional purposes without much success. And Facebook and YouTube for scientists already exist; they are called Facebook and YouTube.

What does the future hold?

Over the last decade and a half, STM publishers have focused on price optimisation and market expansion. Both these strategies have been very successful but there is not much more market to sell into and library budgets are not growing. This leaves publishers with two options for revenue growth: selling to different people (not the library) or creating new products for the increasingly-crowded library market.

This is why we are hearing so much about workflow integration, new products that focus on individuals and departments, and the launch of OA journals that derive revenue from sources other than the library.

This plateau in library funding is to my mind the best thing that has happened in STM publishing in a decade as it forces publishers to start to build user-centric products that appeal to end users, academic departments, funding agencies, and other stakeholders. I am a big fan of academic libraries but focusing exclusively on that market has lead to a paucity of innovation in STM information products. A more heterogeneous market space should lead to more experimentation and novelty – with semantics as the underlying architecture.

Interview by Siân Harris