Carving a place in the enterprise for Semantic Technology by getting past the semantics of â€˜semanticâ€™
Semantic Technologies have much to offer today’s successful business, with regulatory, operational and economic forces combining to require that timely and accurate data from across the enterprise be available on demand and at the point of need. Clear benefits are often disguised, though, by obscure language, serious misconceptions about what ‘the Semantic Web’ could or should be, and an unfortunate tendency to advocate ‘semantic technology’ per se rather than specific solutions to tangible business problems.
The past year or two have seen real growth in the number of boutique consultancies willing and able to help their clients realise the benefits of semantic technology to their workflow, and at the same time there have been advances in the supply of commercial software solutions in which semantic technologies play a significant part. Mainstays of the enterprise software market such as Oracle are embracing these benefits within their core products, relative newcomers like Siri, AdaptiveBlue and TopQuadrant are seizing opportunities to create new markets, and established software companies such as Talis and OpenLink are re-imagining themselves and their proposition in light of new opportunities powered by semantic technology.
‘Semantic’ is not a market
As with many emergent technologies, proponents of Semantic Technology have a tendency to focus upon the technological minutiae of their approach, bombarding all who will listen with talk of ‘ontologies,’ ‘triples’ and ‘reasoning’ instead of emphasising the direct and tangible benefits to a particular market or problem. For semantic technologies to see significant further adoption, their evangelists need to tell stories that relate far more to cost, efficiency, scalability and competitive advantage. SilverCreek Systems, for one, is not winning custom because it uses RDF, a specification at the heart of the Semantic Web. The company is beating larger incumbents in the profitable Master Data Management (MDM) market because its customers are able to see ways in which the use of Semantic Technologies results in a product that is perceived to be better, cheaper, or more flexible than the competition. It is here that most use of Semantic Technologies will stand or fall; freely considered alongside very different technological approaches to solving the same problems and chosen because they better meet the business requirement, not because they are ‘semantic.’
The Semantic Web is not a different Web
Ever since a frequently cited article appeared in Scientific American back in 2001, confusion has reigned as to how Sir Tim Berners-Lee’s vision of a ‘Semantic’ or Data Web might relate to the Document Web through which we are accustomed to navigate today. Despite some false starts, it has become increasingly clear that Berners-Lee and his peers never intended the creation of a new and separate Web created and reasoned upon by software. Rather, they anticipated the gradual enrichment of today’s documents through the addition of structure; structure that whilst largely invisible to the human reader will enable additional value to be easily extracted. At the simplest level, this might be the explicit machine-readable identification of the title and creator of this article, or the highly effective parsing of booking emails from various travel sites that TripIt performs in building unified and actionable itineraries. The bulk of this structure is already there for the taking, with document titles and authors straightforwardly captured during document creation, and TripIt doing the hard work to parse booking emails from airlines, hotel chains, car rental agencies and more; it is easier for a traveller to simply forward a booking email to TripIt than to manually enter travel details into their online calendar. TripIt is then able to automatically populate the calendar, and do more, demonstrating that early scare stories in which the Semantic Web’s potential beneficiaries would need to ‘write RDF’ are simply untrue.
More complex relationships between large data sets are also being illustrated by the growing Linked Data community, which is converting resources such as Wikipedia’s fact boxes, FOAF-formatted personal contact details and the US Census to build a growing pool of explicitly citable assertions of fact upon which any application may draw via that fundamental building block of the Web, HTTP.
Established content owners such as media giant Thomson Reuters are recognising the opportunity to exploit their existing assets in new ways, gradually adjusting the manner in which they are perceived, exploring new business models and placing themselves firmly at the heart of an extremely interconnected information market. With the Calais web service, Thomson Reuters are using their existing in-depth knowledge of people, businesses, places etc to parse and semantically enrich any unstructured text submitted to them. A document referring to ‘Sun Microsystems,’ for example, would be returned to the user along with additional structured information about various forms of the company name, stock data, officers and more, all expressed in a manner that encourages further linking and reuse to continually grow the available pool of data.
Evolving attitudes to data
Proponents of the Semantic Web recognise that data become more valuable when linked to other pieces of data. The monthly sales figures mean far more when studied in the context of retail indices for the market in question. A patient’s vital signs are more effectively interpreted with knowledge of their medical history and access to data on the ways in which similar patients have responded to particular treatment regimes in the past. The rapid fluctuations of global markets are understood by reference to everything from the price of oil to the weather and off-hand remarks from our politicians or business leaders. In almost every area of endeavour, ready access to internal and external information makes it easier for us to reach informed, timely and appropriate decisions; the trick is to ensure that information is presented quickly and in a form we are able to use, rather than becoming overwhelmed by interpreting, processing, and resolving conflicting messages from the flood.
Increasingly widespread adoption of Business Intelligence dashboards within the traditional Enterprise is an attempt to tame the flow of data through the organisation, making it far easier for management to quickly extract intelligence from disparate pools of data in order to respond in a timely manner. No Enterprise exists in a vacuum, however, and far more effective means of integrating externally sourced data into the next generation of these tools is clearly one of the first areas in which the Linked Data principles exploited by Thomson Reuters and others will reach a far wider business audience.
Growing public awareness of data ownership and portability is today dominated by consideration of Facebook, Google and their peers. However, the wider issues that these sites surface are equally applicable to much of the data that modern organisations create, and upon which they increasingly depend. As we respond to initiatives such as those for Data Portability, Linked Data and Open Data, and as we take on board the growing interest in licenses explicitly crafted to promote the sharing and dissemination of data, we are gradually moving toward a realisation that the blanket protections applied to the entire contents of our data centres are perhaps no longer helpful.
Government may be driven by political imperatives in its recently accelerated moves toward transparency, but both private and public companies are also coming to recognise the political expedience of being a little less secretive; and the potentially far more significant business advantages of finding better ways to integrate the Real Time flow — both internal and external — into their decision making.
For all the talk of taxonomies, ontologies, triples and the rest, at the heart of the Semantic Web idea lie simple, pragmatic, principles and technologies that are all about making existing investments in data achieve a greater return. For all those companies that want their assets to perform better, for all those who need to understand their next opportunity before the competition, for all those able to differentiate their ‘core’ data from that which is merely ‘context;’ maybe now is the time to give Semantic Web technologies a fresh look?