A Triptych Software Development Paradigm - Correct System Design: Recent Insights and Advances - page 28

geographic information system+demographic information system systems to de-
cision support software. Perhaps a fourth result of this paper can be claimed: (iv)
Showing, as we believe it does, the structural main parts of a proper presentation
of software.
The current paper primarily presents data models. They are in the style used
in denotational and in algebraic semantics domain, respectively sort denitions.
But we sketch some observer functions and some axioms. The notation used is
that of RSL [1], the Raise Method's [2] Specication Language.
This paper is a torso: It sketches the application of a formal specication
and renement-based software development paradigm to a eld either not very
well understood or covered by AI (articial intelligence) researchers and AI pro-
grammers. AI contributions, valuable as they may be, usually, as do most con-
tributions in software engineering, zooms in on a narrow problem, solvable (ie.
expressible, programmable) in some AI{language (or other). But usually such
contributions do not try to isolate the domain from possible requirements; nor
the requirements from the implementation. Instead the solution \drives" the
development and its presentation.
We advocate a far more comprehensive approach. First we cover, in isolation,
domain problems. In the jargon of software engineering these are the \up-stream"
issues based on whose precise understanding one may formulate requirements.
Then we relate domains to requirements, and nally to software (more precisely
software architectures). Throughout we try to make more precise such software
engineering and such international aid organisation jargon as decision support
system, respectively sustainable development | as well as many other terms:
indicator, equity, etc.
The triptych paradigm: from domains via requirements to software (descrip-
tions, has been covered more comprehensively in other papers, for example the
recent [3,4,5,6,7,8,9], and is claimed to have stood its rst tests of usefulness by
the well-reported work of the last seven years at UNU/IIST 1 . The present paper,
to be fully appreciated by readers not familiar with formal development in the
styles of VDM and RAISE, must therefore be complemented by the study of for
example [10] or [2], preferably both. Then a professional reader can see how to
turn the sketch of this paper into full reality.
We leave it to others to compare the present approach to those of UML etc.

2 Summary Review

2.1 The Application

We Domain analyse (Section 4) the notions of Development as based on Re-

sources . We then analyze the concept of Resources, their Attributes and Attribute

Values and Indicators . Based on Value Indicators we dene Equities . Sustainable

Development is then dened in terms of Equities..

1 United Nations University International Institute for Software Technology, P.O.Box
3058, Macau: http://www.unuiist.iist.unu.edu