Thursday, 27 September 2012

What Engineers Know and How they Know It

W.G. Vincenti was an engineering researcher who wrote a book called "What Engineers Know and How they Know It", which I am trying to get my colleagues at the University to read, to understand how engineering research and engineering practice should fit together.

Vincenti's basic idea is that engineering is not applied science - engineering has six categories of knowledge of which the last five are proper subjects for engineering research, as follows:

1.Fundamental Design Concepts

These are not scientific fundamentals, but instead the design engineer's axioms - a common idea of what the thing being designed is for, its operating principle, and its normal configuration.

2.Criteria and Specifications

Engineers may design artefacts to meet a need defined by others in non-technical terms, but in order to do so they need to transform the general, qualitative specification into concrete, quantifiable performance characteristics.

3.Theoretical tools

The theoretical tools of the engineer may be based in mathematics, science, or be peculiar to engineering. They provide ways of thinking about and analysing design problems.

3a. Mathematical Methods and Theories

The mathematical tools least peculiar to the engineer may be based in pure mathematics, or sciences, but they have been simplified for application to a particular situation by intoducing a set of approximations and assumptions which apply to only that specific set of circumstances. More particular still are the phenomenological theories which practitioners share about things too complex for scientific analysis, even if they have little scientific standing. At the far end of the spectrum are commonly held approaches to design of specific systems, used only because they seem to work, and no better method is known.

3b. Intellectual Concepts

Engineers are less like philosophers than they are like scientists. They are not fussy about where they get their ways of thinking about a design problem from - anything which works is good.

4.Quantitative data

Engineers need physical data to design things. They need descriptive knowledge, of how things are. They need prescriptive knowledge, of how things should be to ensure that the designed item meets the specified need. 

5.Practical Considerations

One can have perfect knowledge in all previous categories and still be unable to design an artefact that works. One also needs know-how, usually obtained from long practice in the profession, and interaction with those who produce, commission and operate the artefact.
6. Design Instrumentalities

Or less opaquely, structured procedures for going about the design of an artefact, ways of thinking about design problems, and judgemental skills. Some of these can be taught directly, but professional competence in these areas comes only from practice.

Vincenti then differentiates between seven ways in which engineering knowledge is generated (I know, he clearly had a lot of time on his hands. I am going somewhere with this....)

1. Transfer from science
2. Invention
3. Theoretical engineering research
4. Experimental engineering research
5. Design practice
6. Production
7. Direct trial

He draws the following table to illustrate the interaction between knowledge types and knowledge generating activities:

Table 7.1 Summary of Knowledge Categories and Generating Activities

So engineering research can generate new engineering knowledge in all categories other than practical considerations. Researchers know about, practitioners know how. This is where I come in.

No comments: