notes-misc-reverseDesign

Currently in society the received wisdom is that you have to start with a goal and then determine what to do by what is necessary for the goal. Examples:

• the derogatory phrase "a product looking for a market"; instead "listen to the customer"
• the derogatory phrase "to a man who has a hammer, everything looks like a nail"; instead figure out what the job requires, then "choose the right tool for the job"
• the idea that "designing" something means starting with the requirements, then using some engineering methodology to calculate how the object should be
• the idea that progress in mathematical science or similar is made by carefully defining what you mean, then performing inferences with those definitions to proceed to conclusions

Let's call this goal-directed design. Call the opposite "reverse design". Taken to an extreme, reverse design sounds ridiculous: imagine some sort of object that looks cool and different (for example, a baton mostly covered with globular incandescent lightbulbs). Now try and find something that would actually be good for.

However, if you look at how truly new things are actually designed, i bet you'd see an interplay between goal-directed design, and "reverse design". I bet you'd find that designers do think a lot about requirements and are guided by them, but also think of new ideas that just seem "cool", either from scratch, or more often, inspired by something similar in some other domain, and they think that's a cool idea, and they wonder how it might be used. I bet often they just hold onto the idea in case they run into some situation in the future that calls for it. Perhaps most such ideas never find a use, but some of them do. I bet many of the more novel inventions didn't start by 'listening to the customer', but rather started as a 'product looking for a market'.

Something i do know is that when people work on applied mathematics, they often don't start by carefully defining everything and then proceeding to the conclusion. Not that they don't think about what they want to say and how to define it, they do. But they also make up assumptions that seem tractable, even if they don't seem exactly right or applicable. Then they see if these tractable-seeming assumptions lead to interesting conclusions or correct predictions. If so, then they work on fixing up the tractable-but-crazy assumptions to something that not so crazy.