Basic, long-standing, core concepts.

What emerged has certainly stood the test of time. Big time. Opinions differ widely on just what constitutes the core constructs. Different people have used different names for these terms, and, although the idea was to specify something formal, the definitions of these constructs were generally sloppy. But here is a reasonable specification, in its most rudimentary form:

There are objects (which might also be called entities, things, or concepts). Objects have unique names.

Objects are interrelated by attributes (which might also be called relationships or properties). Attributes are directional, and they have names.

In other words, things in the world can be represented as a simple directed graph. We could say that there are objects called Chickens that have an attribute called Are. The value of this might be an object called Birds. Birds might have an attribute called Lives-In, which links Birds to the object Barnyard. There might be an object called Mr. Fried, which has an attribute called IS, which connects Mr. Fried to the object Chickens.

There are many popular various of this basic idea that have emerged, and they tend to be of the following nature:

One idea is to make a sharp distinction between the notion of a subtype (or sub-kind or subset) and other attributes. So, our attribute Are might become a core concept itself, and we might name it Is-A. Chickens IS-A Birds, People IS-A Biped, etc. Other attributes like Lives-In would be considered inherently different from Is-A.

We could introduce another generalization. A general term for attributes Lives-In and other similar attributes might be Has-A. In fact, we could stop using special words for attributes in general, and just use the terms Is-A and Has-A. We would then say that Marriages Has-A Wife, as well as a Husband, as well as a Date.

These general ideas are actually old, and actually significantly predate computing. We have been struggling with the problem of describing real world objects (like Cows), real world concepts (like Marriages and Respects), and their interrelationships and categories since the emergence of the earliest philosophers. Aristotle distinguished between objects and their attributes, and carefully studied and described many animals and plants.

What does it all mean for the new Web?

So, what does all this mean to us, today, and what does it have to do with modern Web technology? Well, first of all, these concepts of objects and attributes have spread throughout all of computer science.

There have been some significant extensions, like distinguishing between an attribute that we might call a relationship, which interconnects complex objects or notions (like a driver owning a car) and attributes that interconnect complex objects and notions with atomic or simple things (like a car having a color or a driver having a name). Generally, these latter, simple kinds of attributes are now what we call attributes, and are considered inherently different from (and simpler than) relationships.

Another extension that has become a core concept in programming languages is something we might call an object identifier, which is a unique number or other identifier for individual objects; this allows us to carefully distinguish between two people who have the same mother, and two people who have mothers who just happen to have the same name.

Programing languages also introduced the concept of methods, or little programs that can give life to objects. You might be able to tell a marriage object to tell us the names of the husband and wife.

But basic concepts have not changed. There seems to be something natural and fundamental about them.

Building a new world out of old concepts.

And the Web? A revolution is happening today. We are developing languages that allow Web designers to embed machine-readable specifications in Web-resident information. This will largely automate the process of searching the Web, as well as the integration of information at multiple sites. This will in turn lead to the discovery of knowledge by putting together diverse information from across the Web. We have discussed these emerging technologies in the previous postings of this blog; they are heavily and deliberately built on top of ideas that date back to the 1950’s, and in fact can trace their roots to ancient Greece.

DBpedia.

It’s called DBpedia. A former graduate student at my university, Greg Ziebold, pointed me toward it. The goal of the DBpedia is to transform data from the Wikipedia into a chunk of the Semantic Web. To do this, DBpedia is using RDF technology, something we have discussed is past postings of this blog. Behind RDF is an extremely simple concept, but one that has proven extremely powerful and versatile.

The general idea is to break knowledge up into “triples” that describe relationships between pieces of information. These triples can be chained together to discover new relationships. And, importantly, triples must make use of widely shared sets of terminology, called namespaces, in order for knowledge from different places on the Web to be properly chained together.

RDF, triples, assertions, and inferences.

A thorough example can be found in a previous posting of this blog.

Here is a very simple example of triples (also known as “assertions”) and how they can be put together into “inferences”.

Assertion 1: Joe is tall.
Assertion 2: Tall People should try out for Basketball.
A new inference: Joe should try out for Basketball.

Keep in mind that we would want to make sure that the words used in these assertions have precise, global meanings. We might take the terms in these two assertions from a basketball namespace, one that would carefully dictate exactly what “tall” means in the basketball world. Certainly, it would be quite different from the meaning of “tall” in a kindergarten namespace.

More on DBpedia.

There’s a fancy word for sets of triples that use namespaces and represent various areas of knowledge. They are called “ontologies”, taken from the term used by philosophers to argue about the existence of various things, like God. The DBpedia is essentially a vast ontology, formed from triples and namespaces. Most of the knowledge defined by this ontology comes from the Wikipedia. The folks behind the DBpedia have been given direct access to the flow of information into the Wikipedia, so that the DBpedia can stay current.

One way to look at the DBpedia is that it takes the Wikipedia and reforms it into something that can be searched far more effectively. Right now, to search the Wikipedia, most of us simply type in terms (either into Google/Yahoo or into the Wikipedia search page). We try various terms and follow links inside the Wikipedia until we find what we think we are looking for. With the DBpedia, users can search with SPARQL, a language based on the structure of SQL and engineered specifically for searching large bases of triples. SPARQL allows us to traverse networks that consists of triples linked by inferences.

That way, if we were a coach looking for promising candidates for our team, we would use SPARQL to make the connection between Joe being tall and the fact that tall people should try out for basketball. This is clearly much faster and more accurate than googling things like “tall”, “basketball”, etc, until we happened to find Joe in one of the web pages that pop up.

The DBpedia website, by the way, claims to have a triple base that consists of 274 million RDF triples.

More on this in the next posting.

Let’s step back for a moment.

Just how universal is this notion of RDF-style triples? Will we ever have something substantially more useful, more powerful in the semantics it can express?

Data, Information, Knowledge, and Ontologies.

Academic and industrial researchers in computing like to trivialize big words. Let’s briefly look at the problem. “Data” is an old word, and most of us have a sense that virtually anything stored digitally can be considered data. This includes applications and other pieces of software, too. If you back up some applications to free up space on your hard drive, you’ve just turned applications into data, right?

“Information” is a word that came into play when researchers wanted something that was smarter than data. The word was broader, and vaguer, but information was essentially data that was ready to be used by interactive users. If I pull down a page from the Encyclopedia Britannica site, it’s filled with information.

Then, there were demands for an even richer word, one that suggests data that is beyond information, stuff that is rich in semantics that can be easily extracted. Often, knowledge was data or information that had been interconnected, turned into trees or graphs. Traversing the links in the structure told us how various things were interrelated and thereby exposing powerful semantics. The Web in a sense is knowledge. I can follow links between pages to discover how various pages on the Web are interrelated. I can follow connections on the Britannica site to connect a scientific discovery to the story of the discoverer’s life.

Here’s something significant. This blog and all its postings are related to new web technology, such as the Semantic Web. Our central concern has been the partial automation of the searching of the Web, so that users aren’t limited to typing words into Google and getting back stuff no richer than pages that happen to have these words in them. As it turns out, the term “knowledge” dates way back before the days of the Web, but back then, our notion of what it meant to be knowledge and not just data or information was pretty much the same as it is now. Knowledge can be processed by programs, thereby automating the task of finding the right knowledge and applying it to our problem domain.

Then came “ontology”. This is a relatively new word, but it’s perhaps the most embarrassing. The word, until recently, was reserved for philosophers to use. An ontological argument is an argument about the existence of something. Over the centuries, one common subject of ontological discussions has been the existence of God.

Hmm.

The same old, same old.

Flash forward to the Internet age: Computer researchers use the term to refer to a precise specification of the objects and properties (of these objects) in some well studied domain. I guess the idea is to suggest that we can capture the true nature of the existence of some domain.

These domains could be large, like banking, health insurance, or the stock market. Laying out all of the objects involved in one of these is a daunting task. Consider an insurance claim and all of its properties: type of claim, provider of medical service, patient name, etc., and then imagine laying this all out for insurance policies, underwriting tables, actuarial data, etc. To include all of the objects and properties involved in building software for an insurance company would lead us to thousands of interconnected terms. Triples, in other words.

Or our ontology could be the specification of a pencil object, which has properties like being made of wood and graphite and metal, of having yellow paint and a little pink eraser. Triples like this:

The pencil has a pink eraser.
The pencil is painted yellow.

This characterizes the nature of the challenge we have taken on in our efforts to build ontologies. We take on the problems of scale, not the problems involved in really capturing, in some formal fashion, the nature of the world around us. We build gigantic, but very simple, models of the things that concern us in the software world.

We have trivialized this term, ontology. In fact, for the most part, we’re simply referring to the same old, same old modeling construct: triples. Yes, that simple tool called RDF can be used to build a vast “ontology”.

There is something about the nature of triples that has conquered computing. It is a concept that, as we have seen in previous postings of this blog, underlies object-oriented data structures. It predates object-oriented languages, going back to the early days of AI and the attempts to model the real world.

So, what is an ontology?

An ontology is supposed to be the end of the Semantic Web rainbow: our ability to fully automate the specification and searching of the real world. But the next time some computer person tries to impress you by tossing this term at you, remember to just shake your head and say “Quit being a puff toad. You’re just talking about triples.”