Zygmunt Ryznar

SOME REMARKS ON "REAL-WORLD" INFORMATION SYSTEMS

 "The basic challenge to the managers of today is to foresee and manage
the flood of changes facing their organizations" (D.H.Sanders)

Introductory remarks

There are several views that could enable or trigger a radical progress in information systems (IS) development.

First, IS should reflect (at least understand) a real-world in terms of events, processes and objects. To reach a capability to cover a process flow one must reorientate from a static data-view to dynamic view containing new objects like spiral and temporal state(s) objects.

Secondly, more friendly interface to the user is required. It implies a specification language oriented mostly on business (i.e. real world, way of thinking of executives) and equipped with imbeded properties to convert into information technology ( IT ) objects.

Thirdly, there must be changes in core of IT to enable a freely structured system, which keeps all temporary links and transformations of objects.

 

Features of the real-world information systems

......A reality shows us a complexity and a changeability. Events flow and objects interact with each other. To meet these requirements IS should be pervasive and flexible. The best flexibility might be assumed in systems with variable structure having changeable relations and borders.

No

data-view

real world wiew

1

information

knowledge and data-mining

2

data specification

problem specification

3

static view

dynamic view

4

sequential processing

pervasive on-line computing

5

files processing

event triggered processing

6

a' priori defined programs

dynamically assembled programs

7

a'priori defined reports

dynamically composed reports

8

...........

.....................

9

...........

.....................

Structuring

Real-world seems to be ill-structured (at least non-hierarchically). A critical view on the hierarchic approach one may find in [LAN-73]: "..commonly quoted rule (...) is the use of an hierachic structure. This is often suggested on the ground that natural systems are often (or even always) built upon a hierachical structure. Again this is not a very satisfactory state of affairs, by not providing an answer to the question why and does not give a sufficient insight. Also although a set of rules or principles is fairly generally accepted, not only does the literature lack good arguments for their support, but also an explicit listing of them is still missing and the rules are seldom operatively precise enough".
In this context I must here confess in weakness of my arguments, which are rather intuitively formulated.
I argue for "freely structured" approach in following cases:
  1. in multi-purpose or multigoal systems with common components (where the same components migh be placed at different hierarchical levels or even in a separate trees)
  2. when it is necessary to distinguish between a physical relation "part of" and a logical relation "subordinated to" (they will be located probably in different hierarchical "paths")
  3. when a decomposition cannot be made because do not exist clear and exclusive criteria
  4. when do exist temporary (or temporarily defined) relations that depend on future events.
An inspiration to freely structuring approach might be a contingency theory, chaos theory and events theory.

 

[LAN-73] Langefors B. Theoretical analysis of information. Studentliteratur. Lund und Auerbach 1973