To begin: Semantics is the study of meaning. It focuses on the relation between words and phrases and due to the relationship what they mean in context. Semantics is a game changer when it comes to the management of large volumes of structured and unstructured data. Having the ability to understand words in context using lexical or linguistic semantics will drive linked data initiatives and ultimately, as many believe, will make Web 3.0 a reality. With this in mind, possessing a “semantic platform” is the new marketing target for many software companies and they are quickly moving to describe themselves as a semantic platform in hopes of aligning themselves with the semantic snowball as it continues to gather speed.

Lexical semantics is a linguistic theory that investigates the meaning of words in context. It is the study of meaning, as it applies to words, phrases, sentences and at times larger units of information along with and how humans express themselves through the use of language. This theory understands that the meaning of a word is fully reflected by its context. Therefore the meaning of a word is discovered by its contextual relations with others words.

The idea of semantics is expressed in many forms that include the semantics of programming languages, formal logics, and semiotics. Additionally, the term semantics is used to refer to different types of data structures that have been specifically designed and used for representing information.

Ironically, due to the variety of ways the term semantics is used is why such a broad range of software companies are able to state that they have a semantic platform.

A semantic network is a network which has been built (generally by Linguists) to represent the semantic relations among words to derive meaning in context. This network is most often used as a form of knowledge representation. An example for Mammal and several potential relationships is represented below.

A semantic platform is a software infrastructure that is able to pull in undefined data and push out defined data with the proper meaning attached in the form of new semantically relevant metadata describing the unstructured content. This in turn creates new insight and knowledge for people to access at the right time in a usable presentation.

Deploying a semantic platform allows companies with the ability to drive a variety of corporate initiatives around knowledge management with the end goal of providing the right information to the right people at the right time. If organizations are interested in making their knowledge more easily found and more usable, deploying a semantic platform is the route to pursue.

Once the decision to investigate what semantics has to offer is where the confusion will often begin.

Over the past 9 months I have attended 6 trade events. Each event was well attended, well organized and professional. The attendees at the events came from a variety of companies that ranged from oil & gas, media & publishing, government intelligence agencies, manufacturers and the pharmaceutical industry to name a few.

The companies displaying at the trade show portion of the events varied as well. There were companies that offered approaches with the traditional key word and statistical technologies, ontology companies and structured data platforms, in addition to companies discussing semantic technology. What I found interesting and discuss with great irony is the variety of ways the companies were each referring to themselves as having a semantic platform. Key word and statistical technologies who do not offer a semantic engine have begun delivering semantics as a key marketing message. They are now “training” their sales people to deliver a corporate message that they offer a semantic platform.

Needless to say, this has caused a great deal of confusion for the attendees/customers. A few of the commonalities I have discovered are:

1. The attendees want to learn how they can better discover, manage and organize the volumes of data that their organization is creating and absorbing every day.
2. Many attendees did not have a clear idea on “What is semantics”.
3. Booth staffers used the term semantics loosely and in a variety of ways.
4. Presentations and individual definitions of a “semantic platform” created an additional layer of confusion for the customers to wade through.

To help clear some confusion and weed through the marketing clutter, I have listed several questions around three common topics that every customer should be asking the vendor when looking to purchase a linguistic semantic platform:

Semantic network:

1. Show me how your semantic network works? How does you software manage the ambiguity inherent in language?
2. Explain to me how I can modify the semantic network to reflect the ways my industry / company uses terms?
3. Do you offer different semantic networks for different industry verticals?

Taxonomy:

1. How do I create a category and write a rule to define the type of information that is classified to this category.
2. How do I change the rule/category if I am not pleased with the precision or recall?
3. How many nodes and levels can I create in the taxonomy?

Content Aggregation:

1. How do I create spiders, crawl unstructured data sources and identify the data I want to collect?
2. Can you collect data real time? Show me an example of this?
3. Is faceted navigation a standard option?

Data output:

1. Upon index, is data represented as XML, RDF or both?

As semantics continues to grow and gain momentum in the enterprise, many more companies will work to create new marketing material that will align their offerings with a semantic message.

Please make sure to ask the hard questions up front. If the company is not able to easily answer the above questions, they will be answering the most important question… they do not have a semantic platform that will deliver the value and provide the answers you seek.