Want to keep learning?

This content is taken from the University of Strathclyde's online course, Understanding Information and Technology Today. Join the course to learn more.
crowd of question marks

Ask Viktor

Throughout this course, you will have the opportunity to ask Dr Dörfler the questions that matter to you.

Please post your questions for this week in the comments section below.

This week there have only been 4 comments/questions in the ‘Ask Viktor’ section of the week, this is going to be a relatively short topic this time. However, this is understandable, as week 3 was the ‘slide rule’, so it is supposed to be relatively straightforward, aimed primarily at acquiring a few concepts and learn about some tools and trying them out. However, there was some interesting discussion going on during the week, which I wanted to call everyone’s attention to.

Newer and older tools

It often happens in these nontechnical courses on technology that there are a few techno-savvy people also taking it, and if they are active, the co-creation of the teaching material can get to a whole new level. This time it was Joanne Thoms and Karen Saxl who started off this stream of discussion. The essence of Joanne’s position was that the DFD, ERD, STD, and similar specific modelling tools are generally outdated, today most would use Architecture Frameworks. Here is the explanation Joanne provided to Catriona’s comment

When I use the term Architecture frameworks I go beyond modelling the Information architecture (which I know as the operational view) and also look to model Strategic, System, Service, Technical, Physical, Management (of the Architecture) and Acquisition viewpoints. Each view having a multiple purposes from helping to define the need, through to solution and evolution of that solution.

Of course, Joanne is right. However, I still believe that the right thing is to teach DFD, ERD and STD. The reason is exactly that they are simple, more specifically, that they provide a model of the IS/ICT from one specific well-defined perspective. I also directed you towards the UML, which is a unified symbol system covering the here introduced and other methods as well. This is in a way the first level of synthesis, at the level of symbols. There are then comprehensive frameworks, such as the one Joanna mentioned, which can build models from a variety of perspectives, find relationships between these models, think through relationships with other organisational functions, and so forth. Starting from such a complex standpoint can be very useful for those who get into the technical side of IS/ICT (as vendors, maintainers, etc.), as they will get an understanding of the technical ‘big picture’ before getting into the details. However, for non-technical people, the users of technology, the ‘big picture’ should not be a technical one but one in which technology simply serves a purpose. Furthermore, if needed, the frameworks can be more easily understood if some basic dimensions of modelling are well established.

A related issue was raised by Karen Saxl, who noted the agile/adaptive developmental processes. These do not directly contradict the ‘more established’ modelling approaches, you may use DFD, ERD and STD in agile. However, it is correct that these have much closer links with the more traditional ways of making IS/ICT, when, a long time ago, we wanted to figure out the IS/ICT that we were designing as if it was to stay like that forever. Gradually, however, we realised that we will need to change them, and much more frequently than we seemed to have assumed – and much more frequently than most developers of those time would have liked it. Thus, gradually faster and more flexible approaches were developed. For me the essence of agile is getting something to the user as soon as possible, that delivers a more or less working version of the feature that we want to get to the user. For this we usually use rapid prototyping, modular building blocks, and anything that can help us get something working out faster. Then, instead of predicting the usage and the future, we adapt the software as we go. None of these is, again, contradictory to DFD, ERD, STD – but we do not really use them at the level of detail and comprehensiveness as with the ‘old-school’ approach. The final part of this query was:

Have you had experience of agile/adaptive development processes? Do these represent evolution? Survival of the fittest or dead-end?

I believe that the agile/adaptive developmental processes were necessary, as the Cathedral-style making of IS/ICT could not work. It was too slow and the product far too often mismatched what was needed. In a sense, this does represent evolution, very well observed. However, we need to see evolution as an ongoing process, not something that was taking place until now and we now arrived at a stable state. So I don’t think that the current approach will stay on forever, it will be replaced by something that fits even better the expectations of the future. This does not mean that it is a dead-end, it is a transitory state. Just like any other state…

In this course I have chosen an approach which is perhaps most common in business, but it is user-cantered, examining both the individual as well as the organisational level. Some participants even noted that their primary interest is at the level of the individual. Of course, as I am a business school teacher, this is for me the most natural standpoint. However, this is not the only or even the primary reason for choosing this particular approach: the reason is that we are always members of some sort of organisations (not necessarily business organisations), and usually the organisations will exercise some control over the IS/ICT, often they are the ones purchasing the IS/ICT, so looking only at the relationship between the individual and the IS/ICT would mean missing an important part of the (big) picture.

Thinking machines

Here is a very short one: John Mcconnachie asked “Do you ever envisage a time in the future where systems, computers will out think humans?”

The short answer is NO! The longer version is what we do in week 5. Perhaps, however, this is where I need to say a few words about my background, although the basic points were covered in my introduction in week 1. So, first of all, I have been leading the development of a knowledge-based expert system shell called ‘Doctus’; knowledge-based systems belong under artificial intelligence (AI). I was never doing the programming of this software, I was telling the programmers what to do. So my job was the big picture, I only needed to know enough about the details to know what can and what cannot be done. However, the programmers I was working with were far better programmers than I ever was (I learned programming as a student of mathematical engineering, but it was not my calling). So we were doing different things. The reason that we started working on knowledge-based expert systems was that we wanted to support decision takers by organising the knowledge of their experts in a transparent way. My research, for about two decades now, has been focused on personal (and more recently trans-personal) knowledge. So, I believe, I have a very strong ground for not believing in a thinking machine – however, it is still just a belief. It is not possible to prove that I am right, it may be possible to prove that I have been wrong, if someone ever built a thinking machine.

Educational system(s)

Finally there were two comments/questions about the educational system. On the one hand, I commend Almamo Danfa for turning this back on me by asking how I would draw such a diagram. The answer to this is not so difficult, as I (as an object) will simply belong to multiple object types – i.e. as a teacher and a student. However, if I get this to the next level, where the question noted multiple stakeholders, such as governments, schools, learners, parents, things become a little more messy. I would also add current and future employers. However, we do not need to design an IS/ICT that handles all these. If I was to give a systemic diagram of the Education System, I would not attempt DFD, ERD or STD – I would either go for the already mentioned Rich Picture approach or use causal mapping or knowledge-based expert system. Which one depends on the purpose of the model. A model is like a caricature: it simplifies reality and it emphasises particular attributes of that reality from a particular perspective.

Share this article:

This article is from the free online course:

Understanding Information and Technology Today

University of Strathclyde