On the surface this post may appear to be a non-sequitor from the majority of my posts. However this is where I was in the 1990s while I was working and where I had the itch to do graduate work. Had things been different, this may have led me to a thesis but fate had other plans.
At that same time I was drawn into Virtual Reality. The first blush interest in VR was the visual representation of things. In the intervening years that began to drift toward the visualization of invisible things such as networks and social constructs. So this post is a throwback to that point in my thinking but linking it to where I want to go forward.
When one does coding, each named semantic unit is either a distinct cyber object or a type, that is a generic (which tbh is itself an object.) So my professional desire was to find some way to connect the corpus of requirements discussions with the language of coding. Often there are one-to-one correspondences between object names and realworld objects (often themselves abstract objects). My naive view is that requirements analysis should be able to pluck out of that corpus linguistic tokens that will be manipulated in the automated system.
I saw a paper titled Language quality in requirements development: tracing communication in the process of information systems development which appeared in the Journal of Information Technology, Sept 2013, vol 28, iss 3, pp198-223 by Christopher Rosenkranz, Marianne Covera Charaf, and Roland Holten of Goethe University,Frankfurt. “We propose the concept of language quality as a suitable means for analyzing the emergence of coherent and meaningful requirements. By applying thge thereby developed dimensions of language quality to a real information systems development, project, we are able to obtain practice-grounded proposition to further evaluate the consequences of different actions on the interaction and commmunication process of stakeholders in requirements development.” whew.
More closely aligning the words used in user communication with names in the code should have the benefit of aiding progream comprehension for programmers with domain knowledge.
Dale Fletter Blog Site 