Thursday, May 10, 2007

Point of Data

(Read all of this, it is a Zen Koan.)

Been thinking about data lately (possibly "in" as well). And information, content, metadata, meaning, knowledge, taxonomies, ontologies.. Ontology. Flashbacks of never-ending philosophical debates, epistemology, Plato's Ideals, empiricism, positivism, all of that. Well not so much of the latter really, I guess I've learnt to recognize mental swamps before I go trekking nowadays.

But data. I think a little clarification could be in order. I'd say it goes like:

Data
Particulars (atoms if you will) of that which compose our impressions. Not the world, but that stuff from which we "get the world".
Information
I just quote Gregory Bateson: "a difference that makes a difference". The part of data we can use, as opposed to "noise" or "void".
Content
A tricky thing. "Composited information with a bound context" perhaps, with varying (often hard to measure) complexity.
Metadata
Let's say "added bits used to externally correlate content". I'm not much of a fan of the word anymore. I may deprecate it in favour of just "context-providing statements". The stuff that turns data we don't get into data we get.

I often just call content and metadata "data in" and "data about" nowadays.

Of Content

Content is perhaps the most widely used and less defined stuff. It's abundant, the substance which we structure (and in so doing "contentifying" it further). It is the composition from which richer meaning can be derived. By it's virtue of having a context. This article is "content". That last statement is information (this was a reification, but I digress). Possibly "metadata". Now it's just getting funny. Anyway, content is stuff which can be molded to gain shapes and shades, color and tone; somewhat "synergetical" effects which may or may not add meaning — more often than not depending on cultural aspects (part of the implicit context).

It is stuff which we leave semi-formal at best, the information we hardly can process by anything less than our own neural networks (brains). Somewhat pessimistically perhaps, information which during interpretation gain illusory qualities (akin to optical illusions) which either enhance or corrupt the bringing of meaning.

I won't get into information theory nor semantical discussions. Neither into a discussion of techniques of how to "fluff" expressions to become more sympathetical, making the receiver prone to interpret them as "meaningful".

To the Point

If you're still reading, this is the part where I intend to make a point. Binding the context by which this becomes meaningful.

This content bears no inherent meaning. It's all in your head. Perchance it may have conveyed a structure which made sense, put your mind into a state of "ah, ok". I really just needed to externalize that first part of categorizing some terms that continue to come up in discussions, and for which I needed some binding interpretations.

Perhaps you detected an "anomaly" if you tried to interpret the terms in a hierarchical fashion. "Data in" and "data of" was my differentiation of Content and Metadata, seemingly "instances of Information". But Data isn't necessarily Information. The thing is, that "thing that makes the difference" is where the "meaning", "semantic relevance", "precise form" comes in, and what that is my current state of mind can only grasp by intuition.

We use Content, and the more fine-grained and "to the point" context-providing statements (Metadata), one as compositions within a context, the other to bind both particulars and contexts by means of relations and characteristics. To have a feel for the difference is important, for it is the key by which to understand why Knowledge Representation is the missing piece in many Information Technology issues today.

And this is where How To Tell Stuff To a Computer, and then The Semantic Web FAQ comes in, as my recommended reading of the week, to get you going. Those sources of information will hopefully make the difference that eluded you here.

That's the point.

(Or is it?)

No comments: