Preface
It has been nearly a decade since the second edition of Semantic Web for the Working Ontologist came out, and we are pleased to now be able to present the third edition. While we are gratified to find that a book about technology is still in demand after such a long time (and the first edition was 12 years ago!), some explanation is in order as to why it took so long for a third edition to be written.
For much of the intervening time, we would be occasionally be asked about a third edition. At first, our answer was that the standards had not progressed enough to warrant a third edition. But after the ratification of the Shapes Constraint Language (SHACL) for RDF, the release of Schema.org, and the settling down of a protocol for sharing data on the web (the Linked Data Protocol), this answer became disingenuous at best. The truth of the matter was that we had both moved on to other projects, and were not feeling energetic about updating the material.
Until one day, it was Fabien Gandon who asked the question. Jim told him that we were both over-committed, and hinted that perhaps what we needed was a third author, to bring a new viewpoint and energy to the project. With a little bit of arm-twisting, Fabien agreed to take on that role. So we welcome Fabien as a new author. Without his energy and initiative, this project would never have happened.
In writing this edition, we realized that the experiences we have gained, both positive and not, in working on real projects for government and industry, as well as in large academic networks, had helped us to develop a more mature understanding of what role the Semantic Web stack really can play in much larger scale projects. Further, we have been seeing the role of semantics on the web not just expand, but become crucial to the modern web ecosystem which increasingly includes artificial intelligence, large-scale E-commerce, and an increasing ubiquity of knowledge graph systems. The new examples and modeling techniques discussed in this edition, are motivated by many of the projects we have been involved with that required bringing together many disparate datasets or providing structure to the extracted information from the vast web of unstructured text, which power so much of the machine-learning-based techniques that are crucial to modern enterprises.
There are a number of innovations in the third edition. The biggest addition is a whole chapter on Linked Data, brought in primarily by Fabien, with an emphasis on the Web Architecture behind the Semantic Web. We also revisited all of the examples, and where necessary, brought them up to date. New versions of CHEBI and QUDT have been released since our second edition, and the Good Relations ontology has been absorbed into the larger Schema.org effort (Chapter 14 in this edition). We have updated the examples from data.gov. In this case, there were changes to how data are published (some data sets that were previously published in RDF no longer are), but also changes to technology (the methods for importing tabular data as RDF that we outlined in earlier editions are now available as web services, so there is no longer any need for data publishers to perform RDF conversions themselves.).
We have updated our modeling advice (Chapters 15 and 16), based on experience working on ontologies in the Semantic Web and informed by new insights brought in by Fabien. We have added a small section about BridgeDB, a simple application of Linked Data principles to life sciences. We have updated Chapter 11 to reflect changes in The AGROVOC vocabulary since the second edition.
Probably the most common inquiry we got about the second edition was a request for the data behind the examples in the book. We got requests to host them in GitHub, so that anyone could download them, but this isn’t a very exciting way to distribute example data for a book like this. A download of the data requires students to install their own semantic database to run the queries in the book.
We are happy to announce that for the third edition, all the datasets are available on the workingontologist.org website, not just for download (which they are), but also all the queries are available as well, in runnable form. That is, you can look up any query in the book, and run it against the data, and get the same answer you see in the book. Furthermore, you can make your own copy of the query and try variations to see how it works. The examples in the book have all come alive.
We