> At the time (mid-1990s), “parametric” design programs -- or those that allow the user to input parameters, then scale and/or tweak them -- were just emerging.
The thesis I'm currently writing is using EAs to "evolve" parts of bicycles (right now lacing patterns for the wheel), and I feel what this quote says is still a problem. Most CAD software now allows some paremeters, like certain lengths or radius can be changed. But if changing the value of such a parameter needs significant changes to something else, the end result is not very good. And trying to generate the model programmatically and read the results of the simulation from an outside program is often harder than it should have to be.
--
I've seen these designs many times, cool to read more about them. Someone at my Uni built an upside-down-bike a while ago: http://imgur.com/b8hEGYr
The thesis I'm currently writing is using EAs to "evolve" parts of bicycles (right now lacing patterns for the wheel), and I feel what this quote says is still a problem. Most CAD software now allows some paremeters, like certain lengths or radius can be changed. But if changing the value of such a parameter needs significant changes to something else, the end result is not very good. And trying to generate the model programmatically and read the results of the simulation from an outside program is often harder than it should have to be.
--
I've seen these designs many times, cool to read more about them. Someone at my Uni built an upside-down-bike a while ago: http://imgur.com/b8hEGYr