display | more...


My field of work and study is 'Information Engineering', a young profession that hasn't really been properly defined yet. Every Information Engineer has his or her own view on what the purpose of this area of work should be.

Personally, I live by this rule: There us no such thing as information overload, there are just vast amounts of data

I submit that the seperation between data and information should be a strong one. Information is no more than data with added value to the recipient, in effect, information is a subset of data. Information Engineering should in my opinion therefore be the structuring of data in a way that as much value is added as possible, and thus generating the max amount of information possible. This WU is basically no more than a treatise of techniques available and request for comments on some ideas of mine.

On data modeling/analyses

Data modeling isn't new, it describes a set of techniques used to describe data, data flow, correlation and structures. Examples of these techniques are process diagrams displayed as UML Activity diagrams and Data Flow diagrams.

The structures allow us to track and analyse the data in question.

Information Analyses.. the next step?

When at one time, I needed a diagram about what information piece X was based on or derived from, I realized that just analysing and modeling data might not have been adequate.

Data modeling structures data so it can be turned into information, bur it doesn't structure information. That is, information is data with added value but data analyses doesn't distinguish regular data from information.

An example: A survey form is sent from a respondent to the surveyor, who then has to check if fields A and B are filled in correctly. Then it's passed to a marketing department who need all data on advertisement appeal.. which are fields C and D. And finally the statistics guy needs the personalia which is field A, B and E.

In data analyses fields A to E would end up in all departments mentioned (it is after all, 1 form which won't be cut into pieces). But only some fields matter to some people, these fields contain information. So when tracking information the separation between information and data does matter.

The problem with information is though, that it can hardly be considered tangible, or easy to say whether some information is valid or not. This should be the second area of focus. It should be possible to point to what a certain piece of information is based on, recursively to a level the analyst deems necessary, Even if that means establishing the earth is round or even more basic levels.

Which leaves us with the last point.. What determines the value of information?

About Information's added value

In my opinion the following items determine the criteria of valuable information:

  1. The recipient
  2. Time of arrival at recipient
  3. Authority of the source
  4. Low effort to absorb the information
  5. The 'resolution' of the information

The first three points can be taken care of: Make sure you get the right information to the right person at the right time. But after that, it gets trickier.

That's because there's a negative correlation between items 4 and 5, If you increase the resolution (add more details) you come to a point where it just takes to many effort to sift through all the information, which then reverts to data again. But if the resolution is too low, the information might obviously become worthless as well.. and revert to data.

So a delicate balance is required. Information Engineering including information analyses should try to find that balance so all criteria are met optimally,