The three principal types of information systems

Information System Classification Scheme

Definitions

First, let us define that there are only 3 basic types of information systems. They are all high value but for different reasons.

Type I systems take a large amount of data, and turn it into a small amount of data.

Examples:  Business intelligence, reporting, analytics, data warehousing.

Characteristics: high I/O, high storage, low compute

These are systems which collect information, often from other information systems, and perform statistical and analytical computation to produce refined information for decision-making.

Type II systems take a small amount of data, and turn it into a large amount of data.

Examples: Simulation, Computer programming, Rules-based systems, Machine Learning, Generative AI, etc.

Characteristics: low I/O, high compute

These are systems that take a small amount of effort or information and transform them into a large amount of information.

Type III systems take some amount of data and turn it into the same amount of data.

Example: Business Process Management, Case Management, Record Management, Document Management, Forms management.

These are systems which take human input and record it for human use. These are the electronic equivalent of a filing cabinet, desk drawer, or file folder. They control access, route data from to person, track changes.

Often, they feed into a Type I component to make sense of all the human entered data as well as human performance data (meta data). Yet, business users are surprised to see that Type III applications do not typically require large computers. That is because they are fundamentally operating at a human rate limited pace and volume.

Why is the storage so low? Because there's only so much data humans can type per day.  Now, uploading videos and photos does increase storage requirements so this is changing.  But the number of uploads is typically fairly limited (at least in corporate applications.)

Characteristics: low I/O, low compute, high flexibility, low storage

Classifying commonly used systems

Now that we have the basic building blocks, we can define information systems by combining these blocks into a diagram that expresses not just information flow, as in a typical box and arrows diagram, but also the performance characteristics.

What type is a spreadsheet?  Depends on the spreadsheet.  They can model all three types.  But they are best at Type I and Type III. 

What about a tracker, like Jira?  It's a type III system - it records information at human speed.  It does not need a lot of computational power.  But it must respect human centered design in being responsive (which is not the same as speed).

What about a word-processor?  When used to edit documents, it is clearly Type III.  When used to perform mail merge, it becomes a Type II.

What about chatgpt or any other large language model? Clearly type I during the learning phase but then turns into type II when being used.