The role of metaphor in computational science and innovation in general

Of late I have been doing a lot of work on innovation. In particular how to explain it, how to foster it and how to manage it. Metaphor plays a key role in the process of explaining and describing innovation and how we foster and manage it. And it is to this use of metaphor that I want to further examine.

When we step back and look around at what is going on in the wider world metaphor is all around us. In the case of human endeavor and creativity we often use the term "the lights going on" as a metaphor for the point at which we understand some complex problem enough for a solution to jump out at us. In the field of human conflict we often use the term "collateral damage" as a metaphor for human death through accidental friendly fire and indeed deaths that arise from poorly targeted munitions that result in civilian deaths and casualties. In the world of science Dorkin uses the term "the selfish gene", not to denote any real act of selfishness but as a metaphor for the process of natural selection. And finally in the world of computer science we habitually use the term "bug" as a metaphor for "defect" and we use the term "bus" to denote some mechanism for the transportation of information.

We use metaphor well when it fits properly to some conceptual model of that which we wish to describe and we use it poorly when we introduce ambiguity and sometime simply it is wrong.

My bug bear is the metaphor of the funnel in innovation. According to wikipedia "A funnel is a pipe with a wide, often conical mouth and a narrow stem. It is used to channel liquid or fine-grained substances into containers with a small opening. Without a funnel, spillage would occur." Consider this carefully. You have 100 ideas (early innovative ideas) and you use a funnel. 100 ideas go it and 100 ideas come out. And yet we habitually use this to denote some process of capture, prioritization and selection of ideas. When we get the wrong metaphor to denote a process which needs management it makes it harder to implement the correct process because it focuses out attention in the wrong way and at the wrong things. It might seem small but getting it right is as important as getting names right in descriptions of software systems [Milner - Wot's in a name].

So what would the correct metaphor be for this innovation mechanism? When we think about what we are doing to a stream of ideas and innovations it is easy to get the right one. What we have is a filter and not a funnel because we are not selecting all ideas and innovations to go forward, we are selecting the ones that are appropriate in some way. If we imagine having the usual array of crazy mechanisms for gathering ideas such as drop boxes, inevitably we will get ideas such as "I want more rice pudding in the canteen" along side some worthy ideas. But we would not for a moment imagine running with "rice pudding" as a creative or innovative idea, hence the funnel is wrong and the filter is right.

BTW if anyone want's a copy of the Miler paper let me know. Alas since Robin passed away his paper is not longer online.