by John Biderman
I will be attending the 2011 Semantic Technology Conference (“SemTech”) in June on my own “vacation” time in order to further my knowledge of what is going on in this very interesting and exciting arena. I went to this conference in 2006 when it was only a few years old and had about 800 attendees. This year there will be around 1,200, I’m told, and it is particularly interesting to see the greatly increased number of case studies and real-world applications as compared to five years ago, when the event was more academic and futuristic – more about the promise than the reality. The reality is here: semantic technologies are burgeoning. It is my belief that all data professionals (modelers, metadata managers, taxonomists, data governance practitioners, information architects, etc.) should stay abreast of these developments and get schooled in the technologies, not just for career development purposes but because successful development of semantically aware systems depends on the disciplines of information analysis and modeling that we’ve been applying in the relational world for decades. In other words, the semantic world needs us.
Semantics, of course, have been implicit in information systems since the very first columnar data set was instantiated – made somewhat more explicit when column headings were introduced, and even more explicit when relationships, cardinalities and constraints were declared in RDBMS metadata. All these things wrapped some real-world meaning around data, but nothing like the formal representation of knowledge domains that happens in ontologies. I am an avowed proponent of enterprise data models (see previous post) for a host of beneficial reasons, but I also believe that enterprise ontologies will become more important for representing the interrelationships and inherent polyhierarchies of enterprise information. In fact, if the ontology is granular enough, the data model can be derived from the ontology. (This is precisely the methodology the firm Semantic Arts inventively follows for designing Service Oriented Architectures: the atomic information that supports a business process is modeled in an ontology, and they extract the XML structures for SOA canonical messages from that model.)
The realm of semantic applications is going to be increasingly extensive, for example in data-integration programs, rule-based systems, analytics (think inferencing for BI). Semantically aware search capability will change the nature of information retrieval in organizations. We will have ontologies sitting on top of relational databases, to allow semantic search to operate on both structured and unstructured information – in other words, to query by meaning rather than by value, and to do so across enterprise information assets.
Semantic technologies are all about the metadata, which in turn is all about vocabulary, relationships, and meaning. Thus the clear relevance of the data professional’s craft in the world of semantic systems. A whole organization concerned with Semantic Data Management has evolved in the last year. A very interesting article by Brian Schulte describes how semantic modeling will have an important role in Master Data Management.
In other words, folks, this stuff is real and it will be coming on like gangbusters over the next several years. If it is mostly new to you, an excellent introduction is Dave McComb’s highly readable book, Semantics in Business Systems (2003, Morgan Kaufmann). (The publisher put it out in their “Savvy Manager’s Guide” series, but don’t let that throw you off.) It traces semantics from the early RDBMS days right through to the Semantic Web. For a more advanced look at the underlying information modeling technologies, check out Semantic Web for the Working Ontologist by Dean Allemang and James Hendler (2008, Morgan Kaufmann). For a purely layman’s prognostication about where the Semantic Web will be taking business and commerce in the next decades, read Pull: The Power of the Semantic Web to Transform Your Business by David Siegel (2009, Portfolio Hardcover).
Better still, consider taking a class in ontology modeling and/or RDF data stores. If you have experience or familiarity with Object Role Modeling (ORM), the move to ontologies will not be that great a leap – it involves the same declarative subject-predicate-object triples.
My message is that we should all get and stay educated about semantic technology. It is going to change our careers while leveraging our fundamental disciplines.
I will be attending the 2011 Semantic Technology Conference (“SemTech”) in June on my own
“vacation” time in order to further my knowledge of what is going on in this very interesting and
exciting arena. I went to this conference in 2006 when it was only a few years old and had about
800 attendees. This year there will be around 1,200, I’m told, and it is particularly interesting
to see the greatly increased number of case studies and real-world applications as compared to five
years ago, when the event was more academic and futuristic – more about the promise than the
reality. The reality is here: semantic technologies are burgeoning. It is my belief that all data
professionals (modelers, metadata managers, taxonomists, data governance practitioners, information
architects, etc.) should stay abreast of these developments and get schooled in the technologies,
not just for career development purposes but because successful development of semantically aware
systems depends on the disciplines of information analysis and modeling that we’ve been applying in
the relational world for decades. In other words, the semantic world needs us.
Semantics, of course, have been implicit in information systems since the very first columnar data
set was instantiated – made somewhat more explicit when column headings were introduced, and even
more explicit when relationships, cardinalities and constraints were declared in RDBMS metadata.
All these things wrapped some real-world meaning around data, but nothing like the formal
representation of knowledge domains that happens in ontologies. I am an avowed proponent of
enterprise data models (see previous post) for a host of beneficial reasons, but I also believe
that enterprise ontologies will become more important for representing the interrelationships and
inherent polyhierarchies of enterprise information. In fact, if the ontology is granular enough,
the data model can be derived from the ontology. (This is precisely the methodology the firm
Semantic Arts inventively follows for designing Service Oriented Architectures: the atomic
information that supports a business process is modeled in an ontology, and they extract the XML
structures for SOA canonical messages from that model.)
The realm of semantic applications is going to be increasingly extensive, for example in data
integration programs, rule-based systems, analytics (think inferencing for BI). Semantically aware
search capability will change the nature of information retrieval in organizations. We will have
ontologies sitting on top of relational databases, to allow semantic search to operate on both
structured and unstructured information – in other words, to query by meaning rather than by value,
and to do so across enterprise information assets.
Semantic technologies are all about the metadata, which in turn is all about vocabulary,
relationships, and meaning. Thus the clear relevance of the data professional’s craft in the world
of semantic systems. A whole organization concerned with Semantic Data Management has evolved in
the last year. A very interesting article by Brian Schulte describes how semantic modeling will
have an important role in Master Data Management.
In other words, folks, this stuff is real and it will be coming on like gangbusters over the next
several years. If it is mostly new to you, an excellent introduction is Dave McComb’s highly
readable book, Semantics in Business Systems (2003, Morgan Kaufmann). (The publisher put it out in
their “Savvy Manager’s Guide” series, but don’t let that throw you off.) It traces semantics from
the early RDBMS days right through to the Semantic Web. For a more advanced look at the underlying
information modeling technologies, check out Semantic Web for the Working Ontologist by Dean
Allemang and James Hendler (2008, Morgan Kaufmann). For a purely layman’s prognostication about
where the Semantic Web will be taking business and commerce in the next decades, read Pull: The
Power of the Semantic Web to Transform Your Business by David Siegel (2009, Portfolio Hardcover).
Better still, consider taking a class in ontology modeling and/or RDF data stores. If you have
experience or familiarity with Object Relationship Modeling, the move to ontologies will not be
that great a leap – it involves the same declarative subject-predicate-object triples.
My message is that we should all get and stay educated about semantic technology. It is going to
change our careers while leveraging our fundamental disciplines.
This entry was posted on May 4, 2011 at 5:19 pm and is filed under Blogs, Data Topics, Discussion, John Biderman, Metadata, Semantic Technology. You can follow any responses to this entry through the RSS 2.0 feed.
