Semantic Web Applications Graham Klyne Nine by Nine http://www.ninebynine.net/ 26 February 2004

Semantic Web Applications Nine by Nine Who am I? Scientific, engineering and networked software systems architecture Motion capture, mechanism design, IP address translation, MIMEsweeper Internet and Web standards Internet fax, email, instant messaging, content negotiation Most recently, Semantic Web (RDF) I believe this technology is set to have a big impact on computer application development 26 Feb 2004 Semantic Web Applications

Data – Knowledge – Meaning Deep philosophical territory: not going here Meaning? Knowledge Data Physical In the limited sense of KR semantics RDF Applications structure character raw XML Unicode Bits, Octets The range of ideas covered by terms like data, information and knowledge and meaning can extend from measurable physical phenomena (e.g. the orientation of a magnetic domain, or the presence or absence of an electrical charge), to a deeply philosophical concept of meaning which may defy clear-cut definition. This presentation is concerned mainly with two intermediate layers: data, which I regard as an arrangement of symbols, and knowledge which has an associated semantics. (The term knowledge is used here in the limited sense as it appears in the sudy of knowledge representation.) Such semantics arer sometimes described in terms of of propositions (statements about the world, or some world) that may be asserted as facts, independently from any particular arrangement of symbols that may be used to represent them. Despite some commentary, I do not see that the Semantic Web is an application of Artificial Intelligence (AI) in its broadest sense, even though it may draw some important ideas from AI research. 26 Feb 2004 Semantic Web Applications

Open Building Blocks Standing on the shoulders of giants Much recent progress in Internet software is built upon open standards and open building blocks Open infrastructure is not hostage to proprietary systems vendors There's a freedom about the Internet: As long as we accept the rules of sending packets around, we can send packets containing anything to anywhere. - Tim Berners-Lee Semantic Web technology follows this pattern . SOAP XML HTML HTTP MIME SMTP TCP/IP . PHP MySQL Apache Perl libwww Linux Sendmail BSD Much of the recent progress in networked computer applications has been built on the foundation of open systembuilding blocks, in the form of open standards and open source software. Modern applications are too large and complex to be self-contained: they incorporate many pieces from external sources. The big advantage of open systems is that they are not controlled by any one organization. Instead, multiple organizations and individuals may collaborate to produce systems for the greater common good. And most importantly, the systems continue to be available and supportable if any one organization should decide that it no longer suits their purposes to support a given standard or piece of software. Thus an organization that builds upon open building blocks can have confidence that their development investment will not be rendered useless by the decisions of another. Thus, there is a kind of safety in open systems building blocks that is not available to the users of proprietary designs. The rapid growth of the Internet and World Wide Web is a testament to the value of building systems in this way. In an environment of open building blocks, the various available components are often designed to work with each other, based on common standards, rather than to compete with each other. The topic of this presentation, Semantic Web technology, is another set of pieces in the growing armoury of open systems building blocks. 26 Feb 2004 Semantic Web Applications

Evolving Use of Information Consider computer applications that should be sharing common information Typically, data is not readily shared Data must be re-entered or converted, which is expensive and error-prone The Semantic Web is many different things to different people. The view I wish to promote is of a technology to enable and facilitate data sharing between applications. Typical computer applications do not readily allow information to be shared between them, which means that if sharing is to be achieved then it must be converted or re-entered for use by different applications. 26 Feb 2004 Semantic Web Applications

Semantic Web Applications Example Employee information distributed across disparate IT systems Employees Payees HR Finance Users As an example of data sharing, consider software packages used by an organization to deal with finances, human resource management and IT network security. All of these involve information about people who work for the information, much of which is likely to be common across applications. But, in general, the information used by such applications tends to be maintained in a private format, inaccessible to others. In this way, the applications usurp ownership of data that should really be available for the users to do with as they see fit. Security 26 Feb 2004 Semantic Web Applications

Semantic Web Technologies Proposed benefits Re-use information designs Use open building blocks to process common information Integrate data sources: new uses for existing data Collect, Process, Extract This slide previews some goals for deployment of Semantic Web technologies: to provide a framework within which Design costs can be reduced by sharing information designs among different applications Use of a common information framework allows much of the processing to be performed by generic software, rather than by bespoke software for every application Data from multiple existing data sources can be integrated, enabling it to be put to uses not envisaged by the original applications 26 Feb 2004 Semantic Web Applications

The Semantic Web Evolving the Web Evolution of the Web to a network of application-usable information open standards from W3C open software from many sources An open-ended framework for combining and exploiting information from a wide range of sources Semantic Web specifications and technology are being developed by the World Wide Web Consortium(W3C), a widely respected grouping of several hundred leading Internet technology companies and headed by the “inventor of the Web”, Tim Berners-Lee. The mission of W3C is “Leading the Web to Its Full Potential…”. (See http://www.w3.org/ for more about W3C.) It is part of Tim Berners-Lee’s original vision for the World Wide Web, creating a global network of information that is accessible to automated processing as well as human browsers. As such, it is an evolution of the current World Wide Web, not a thing apart. At heart, the Semantic Web creates a framework for combining information from a range of sources, which may be both local and globally provided from anywhere in the Web, and deriving new data that itself may be similarly used. 26 Feb 2004 Semantic Web Applications

Semantic Web Building Blocks Resource Description Framework (RDF) An XML-based standard knowledge representation format for exchanging arbitrary information Web Ontology Language (OWL) A standard for describing classes of objects and enabling inference RDF Query, RDF Rules, Access, and more Pre-standardization, software components Currently, Semantic Web technology consists of two core specifications that achieved full W3C Recommendation status (roughly, a Web standard) in February 2004: Resource Description Framework (RDF) http://www.w3.org/RDF/ Web Ontology Language (OWL) http://www.w3.org/2001/sw/WebOnt/ RDF is the base language that is used to describe arbitrary information. It has a single syntax that can be read by any RDF-conformant processor (complementing XML, which provides a syntactic framework that is adapted differently by different applications). OWL provides a basis for describing classes of objects and their characteristics in such a way that permits new information to be deduced from raw RDF data describing such objects. Semantic Web technology also encompasses a number of other, pre-standardization, activities in the form of freely available software and designs to deal generically with query and deduction using RDF data, and mechanisms for access RDF data on the Web. 26 Feb 2004 Semantic Web Applications

(Detour: example data) Example from network configuration, describes features of a user and a computer system RDF/XML (link) Notation 3 (link) Graph (link:PNG) The example data can be found on the web at: The XML format is that recommended for exchanging RDF data between applications: http://www.ninebynine.net/Papers/RDFExample-xml.html The Notation3 format captures the same underlying semantics as the XML format, but uses a syntax that is easier for human users to create, read and discuss: http://www.ninebynine.net/Papers/RDFExample-N3.html The graph form highlights the general form of RDF as describing relationships between things (“resources” and “literals”): http://www.ninebynine.net/Papers/RDFExample.png 26 Feb 2004 Semantic Web Applications

What kinds of application? Diverse, semi-structured information Open-ended: evolving functions and data Examples: Personal information management (Chandler) Social networking (FOAF) Information syndication (RSS,PRISM) Library/museum data (Dublin Core, Harmony) Network security and configuration (SWAD-E) Semantic Web technology has the potential to be applied in almost any application, but it isn’t a magic bullet, and is not necessarily most suitable for all applications. Applications that I feel are most likely to benefit from Semantic Web technologies are those that deal with a diverse range of semi-structured data (much of which may be usefully shared with other applications), and which are expected to evolve to perform new functions with existing and new kinds of information. The examples indicated here are all existing applications of RDF, and a corresponding system or project name is shown in parentheses. This list is far from exhaustive. 26 Feb 2004 Semantic Web Applications

What can be Achieved? Integration of diverse data sources Focus on information needs Generate new knowledge Aggregation, Inference, Query RDF Input data Results Generic software functions This diagram shows how Semantic Web technology addresses the goals raised earlier. If input data can be provided as RDF, the following functions can be performed by generic, off-the-shelf (or off-the-web) software: Aggregation of multiple data sources. Given Multiple RDF datasets from a variety of trusted information sources, aggregation of this into a single RDF dataset is virtually free. (This is not true for, say, XML.) Inference to deduce new information. A generic inference engine together with domain knowledge embedded in application-specific inference rules can create new information from a composition of supplied raw RDF data. Query, to extract desired results from an aggregation of raw and deduced information. 26 Feb 2004 Semantic Web Applications

Aggregation + Inference = New Knowledge Building on the success of XML Common syntactic framework for data representation, supporting use of common tools But, lacking semantics, provides no basis for automatic aggregation of diverse sources RDF: a semantic framework Automatic aggregation (graph merging) Inference from aggregated data sources generates new knowledge Domain knowledge from ontologies and inference rules XML has been a great success in providing a basis for distributed computer applications to exchange data. There are many reasons for this success, some of which are: A standardized basis for exchanging structured information Availability of common tools to perform common processing functions But XML does not, of itself, provide a basis for diverse applications to exchange and share knowledge. To achieve this, the information must be lifted to a form that is common among applications, yet sufficient to represent the diverse applications’ information. RDF is such a form. Having a simple, uniform syntactic structure and well-defined formal semantics, information from diverse applications can be represented and combined without loss. Further, logical frameworks that work with RDF data can combine domain knowledge in the form of ontology descriptions and inference rules to deduce new knowledge from the aggregated data sources. 26 Feb 2004 Semantic Web Applications

Aggregation + Inference: Example Consider three datasets, describing: vehicles’ passenger capacities the capacity of some roads the effect of policy options on vehicle usage Aggregation and inference may yield: passenger transportation capacity of a given road in response to various policy options using existing open software building blocks [Example (link:TBD)] 26 Feb 2004 Semantic Web Applications

Semantic Web Applications What needs to be done? Information design Data-use strategies and inference rules Mechanisms for acquisition of existing data sources Mechanisms for presentation or utilization of the resulting information Semantic Web technology provides a basis for information sharing and performing some functions with generic software, but doesn’t solve all application problems. The list here indicates some of the areas that must be addressed by an application using Semantic Web technology. 26 Feb 2004 Semantic Web Applications

Semantic Web Applications Benefits Greater use of off-the-shelf software reduced development cost and risk Re-use of information designs reduced application design costs; better information sharing between applications Flexibility systems can adapt as requirements evolve Open access to information making possible new applications These are some benefits to be achieved by using Semantic Web technology in suitable applications. I expect the last of these (“Wider access to information used by multiple applications”) to be the greatest long term gain, but the shorter-term benefits are also valuable. 26 Feb 2004 Semantic Web Applications

Recommendation: Low risk approach Focus on information requirements this is unlikely to be wasted effort Start with a limited goal, progress by steps adapting to evolving requirements is an advantage of SW technology; if it can do this for large projects it certainly must be able to do so for early experimental projects Use existing open building blocks These are my recommendations for approaching the development of a possible Semantic Web application. The emphasis here is on minimizing risks until the benefits are clearly visible. Any work performed in the area of addressing information requirements and processing strategies is likely to be useful, whatever technology may be used to implement a system. The more formal aspects of information design using RDF can help to expose problems early in the development process. Using RDF’s flexible and extensible framework, it is possible to start with a small part of a problem, and expand it in stages. Existing information designs are easily adopted and extended. There is a considerable and expanding range of free software that can be used for protoyping Semantic Web systems, some of which is of very highj quality. My own free software, Swish, has been developed in response to problems I have encountered in designing applications in the area of network devioce management, and as an attempt to apply the theoretical benefits of RDF in practical ways. http://www.ninebynine.org/RDFNotes/Swish/Intro.html Commercial packages for more extensive handling of RDF are starting to become available. 26 Feb 2004 Semantic Web Applications

Proposed evaluation steps Decide if this can help your application Identify a well-defined, constrained sub-goal Design an initial information model Prototype data and inference rules Explore some variations Develop simple mechanisms to present existing data as RDF Start with a simple sub-problem, scoped to just a few days work This proposed approach to evaluating a Semantic Web application has some similarities with the Extreme Programming approach to software development. It might be thought of as an application of Rapid Application Development principles to information design. 26 Feb 2004 Semantic Web Applications

Semantic Web Applications Where are we now? Semantic Web is new technology about 10 years after the original WWW Many applications are experimental The goals may be inevitable... Applications working together with users’ information, not owning it drawing background knowledge from the Web less dependence on hand-coded bespoke software … but the particular technology is not Semantic Web is new technology, with the uncertainties that entails. I estimate it is emerging about 10 years after the original World Wide Web, and I believe that its impact on use of the Internet will be comparable. On this basis, I expect to see the start of widespread public awareness of Semantic Web applications to start in 2005. This year (2004), I expect to see widespread recognition of Semantic Web elements within the technical community dealing with Internet applications. Although I believe the eventual goals of the Semantic Web are inevitable, the precise timing and technology are not. But current Semantic Web technologies have a significant head start, a growing developer community, and significant mindshare among developers and users. Current applications (early 2004) are mostly experimental, but the underlying technology and tools are stabilizing to a point that true real-world applications can be deployed. 26 Feb 2004 Semantic Web Applications

Conclusion: Semantic Web Technology Today World Wide Web incremental advance Evolvable approach to information Leverages open software building blocks Builds on diversity creating new knowledge enabling new applications Low-risk adoption strategy by incremental, re-usable steps If you would like to explore further, please contact me at: GK@ninebynine.net 26 Feb 2004 Semantic Web Applications

Semantic Web Applications End Contact information: info@ninebynine.net http://www.ninebynine.net/ References: http://www.ninebynine.net/Papers/ SemanticWebApplications.ppt ~.pdf http://www.w3.org/2001/sw/ http://www.scientificamerican.com/article.cfm? articleID=00048144-10D2-1C70-84A9809EC588EF21&catID=2 http://www.w3.org/rdf 26 Feb 2004 Semantic Web Applications

Nine by Nine What can we offer? We have participated actively in development of RDF core standard Developed open source software for inference and proof-checking in RDF data Design RDF applications including CC/PP, a W3C recommendation Offer help with Semantic Web information design and technology evaluation 26 Feb 2004 Semantic Web Applications

Data - Information - Knowledge 26 Feb 2004 Semantic Web Applications

Semantic Web Applications End ? 26 Feb 2004 Semantic Web Applications