An application is a method of solving a problem or getting a defined task done. (An online definition: "The act of applying as a means; the employment of means to accomplish an end; specific use." http://ardictionary.com/Application/7696)

A software application is computer programming that supports an application.

Today, unfortunately, the phrase semantic applications is used most frequently to describe software that incorporates principles of the Semantic Web, Sir Tim Berners-Lee's most recent major invention. (The most frequently cited description is the Scientific American article, The Semantic Web, by Berners-Lee, Hendler, and Lassila. Formal W3C Semantic Web activities are described at http://www.w3.org/2001/sw/.)

But, as good as the esteemed knight is (And he is very good.), it's really premature to subsume all semantic applications under the Semantic Web. Let's grant that semantic web applications involve the Semantic Web. But let's open up semantic applications to include all applications that explicitly specify meaning as an essential part of how they address a problem.

The technologies of the Semantic Web are -- and will be -- vital in many semantic applications. They are a godsend. Sir Berners-Lee has made it very clear that if we want software applications to talk to each other about content in web resources, we have to add explicit, unambiguous semantic value to that content. Of course, that indirectly helps people find and manage information.

Mills Davis of Project 10x frames the requirement in the following way:

A new approach to knowledge-intensive work is needed that delivers not only the information, but all of the theory and modes of reasoning needed to perform a job or task.

Mills Davis, Semantic Wave 2006: Part 1 - Executive Guide to Billion Dollar Markets, p. 10

Also very sensible ... and also heavily oriented to the Semantic Web.

But the current crop of advanced Semantic Web-based technologies focuses primarily on large-scale, isolated applications. For example, computer ontologies and text-mining/text-analytics applications -- some of substantial scale -- are used to support narrowly defined applications. There's nothing wrong with that, because the benefits are real and, well, that's where the money is: the sale of big-ticket software and services.

However, that Semantic Web-centric perspective can be distracting and limiting. The semantic applications we have seen so far scarcely begin to address the pervasive role of creating, managing, and communicating meaning in our work and lives. Our current habit is to build value and manage organizations through information. But instead of producing value directly, we typically create intermediate information objects that have no intrinsic meaning. We barter those objects -- those unstructured documents and rich-media objects -- instead of understanding how our currency of concepts and instrumental ideas can be used to create greater value.

Moving beyond documents

We need to move beyond bartering documents. Full-text search doesn't help us do that, nor do open tagging and bookmarking services like del.icio.us. You may find what you want faster, but you still have to consume what you find. You take only baby steps toward creation of value -- at an immense cost of time and effort. Document management may make bartering documents more efficient, but it certainly doesn't reduce complexity or get us significantly or measurably closer to doing what we want to do.

Yes, these technologies help us deal with the volume of information, but they are barely keeping pace with the steady increase in that volume. And if we continue framing the problem primarily as dealing with the volume of information, we won't ever solve the problem.

• First of all, we have to avoid concentrating solely on the negative: stemming the tide of powerful external forces (volume of information) ... most of which have no connection to the value we are trying to create. We need to improve our ability to ignore those uncontrollable forces. We need to develop the habits and practices of circumventing the waste and errors generated by those forces.
• Secondly, we need to hammer across to managers, workers, and educators the understanding that words are not concepts. This is probably the most basic aspect of what Bob Glushko has labeled semantic literacy. It's so simple. Go look up any number of English words in a dictionary. Take resource, for example. How many meanings does it have? What special meanings, not listed in the dictionary, does it have for the people you work and play with? One of those meanings comes close to being a distinct concept.

  We all understand this simple distinction, and from our own experience we can give thousands of examples of the problems caused by confusing concepts with words. But we ignore this distinction!

  Work is about producing value. You simply cannot assign value to words, a string of characters that has no meaning except in a very limited context -- and sometimes not even then. That's the first positive step.

• Thirdly, unless we accept that words are not equal to concepts, we can't take the next positive steps toward linking what we know more directly with what we want to do. The explicit value in concepts must be connected much more directly with the services and products we create.

If these ideas seem extraordinarily simple, naive, and obvious, that's because they are. The problem is that most of us are ignoring the obvious. And the cost of doing so is huge. Yet we continue to do so ... stubbornly, repeatedly, almost religiously. We aren't grokking these ideas.

And that's why educating and retraining both managers and knowledge workers is just as important as developing new semantic software applications. In fact, the challenge we face is not so much adoption of semantic software as it is adoption of a mindset and practices that value creation and management of semantic assets.

Sample semantic applications

Ironically, we do devote time and effort to semantic-based problems -- but without accepting the basic truths mentioned above.

We spend a significant amount of time organizing our own resources ... but in isolation, and often with a high degree of frustration. And we are constantly re-inventing systems for describing stuff -- sometimes very formally (as in database schemas), sometimes very informally, and sometimes using traditional models (traditional categorization).

Every domain or profession ultimately devotes a substantial amount of time to organizing its knowledge. Each does so differently, even within the domain. Many invent and, perhaps just as often, reinvent entire systems of organizing their knowledge and attempting to record and communicate it more effectively. But they don't see this as an activity similar to those performed in other domains.

Almost all of us use tools in which precise specification of meaning plays a key role, but those precise vocabularies are walled in -- for example, relational databases and process modeling tools.

Given the number of times we have been exposed to the term semantic web in the past few years (On Google, it has eclipsed the number of instances of Lindsay Lohan, for example.), it's easy to pick out a few obviously "semantic" applications, including computer ontologies, some taxonomies, aspects of text analytics/text mining, and various tools and services based on RDF.

But the semantic aspects of the following applications are also not to be ignored:

• Personal information managers (which, ideally, resemble personal ontologies) and outliners
• Tools for visualization of relationships among concepts and idea (concept mapping, etc.)
• Social Network Analysis (SNA), insofar as it associates ideas with people
• Project management

In fact, if you look closely, you can see that many job roles, activities, and software applications depend heavily on precise specification of the meaning of concepts.

Where do we go from here?

We'll explore some of the ideas and products mentioned here in greater depth.