ShareSemantics’ Blog

One of the main benefit of the Information Technology is the reduction of the impact of imperfect knowledge in all domains of human activities. From an economic point of view this means reducing the costs due to imperfect knowledge about material products, financial products and services. In the social and government realm reducing the imperfectness of knowledge means choosing better representatives and policies for the society and the environment and also strengthen democracy. Also in the scientific and technical realm there is a need to reduce the imperfect knowledge about what has been discovered. To reduce the lack of knowledge we need information available and interoperable over a medium. We can restrict ourselves to talk only about the Web as the medium. What it means information availability seems to be quite obvious, if you don’t ask yourself how. What it means information interoperability is well illustrated by the following quote from D. Engelbart article[1]:

“The purpose of interoperability is to avoid having information islands between which information cannot
flow effectively”.

It is useful also noting what Engelbart adds about the overlooking of the interoperability issue:

“Since we grew up in a paper-based framework, we’ve given little thought about how much exchange and interoperability support we really do have, and how much we depend upon it.”

So the questions are:

1) how information is made available ?
2) how to make information interoperable ?

At this point we need to give a definition of what information is in the practical job of a knowledge worker, information architect or software engineer.  It is a set of one or more statements about a thing, real or abstract. That information to be meaningful to a human agent has to have some syntactic structure following the grammar of one known language. This is enough for a human agent to understand the meaning of the information, what that information stands for in the real world, because the implicit semantics of the elements of the statement, words, can be understood by the human agent. That same information, also the most simple one, cannot be understood by a machine. It cannot understand what is the subject and what is the predicate, above all it cannot understand what any single word in the statement stands for. In order for the machine to understand the meaning of the information a more formal structure must be provided, the semantics of the words must be put explicitly inside. You must define a syntactic structure of  statements like putting the subject first , then the predicate and the object and then add the semantics. Adding the semantics means using well defined terms from a shared vocabulary. Given the definition of information and its possible representations, machine understandable or not, we can return to the first question. Since the Web is our medium of choice we can say information is given inside documents from a server to a client or requester using a specific protocol known as HTTP. So the information availability depends on the document availability. The same information can exist in more than one document. In order to retrieve that information a human agent must have a reference to the document, or its URL. If she doesn’t know the URL then she must give a search engine some keywords to be searched in all documents indexed by the engine with those keywords. But what she gets is not the information but a probability that information she is looking for is inside the documents returned by the search engine. So what she has to do is threefold job:

1) get the documents
2) look inside the documents for the information needed
3) extract that information

If more information is needed she is probably going to repeat the sequence more times. So to use that information the user must do all the interpretation tasks. Now having delineated how the information is made available we look at its interoperability. What we have found is that the software agent, or browser, cannot help our user. This happens with the machine unfriendly kind of information resources, that is almost all the documents in the Web, static or dynamic. All this can be alleviated if we use the machine understandable representation of the information. This way our software agent, made aware of the semantics of the information, can help us in the task of retrieving the information already usable by the same or other software agent specialized in the kind of task that we need in our specific job. So to achieve the information interoperability this must be put in a machine understandable representation. Letting the machines help us in our less intelligent tasks can improve the reduction of the imperfectness of knowledge, its social and economic costs.

[1]Douglas C. Engelbart Knowledge-Domain Interoperability and an Open Hyperdocument System, June 1990

Have you ever met a little dog in front of your door that simply enters your home so easily that you cannot get it out ? You feel you must give it a chance. The same worked for me with Wordpress. It installed itself in my host just asking me for the key (database settings) that I think it would not be savy to not start working with it.