I want to discuss several ways of designing optical systems from scratchand then optimizing them.Of the 5 completely different design methods Iwill be covering,4 are semi-automatic or are completely so and not muchthinking or experience is required on the part of the designer.One of the methods,however,requires a lot from the designer-a solidunderstanding of aberration theory,close attention to what has beenpublished in journals and books,especially from more than 20 years ago,alot of creativity,and finally-a certain playfulness and delight in doingoptical design.For me this is the most fun and productive way of doingoptical design and I will start with a few examples,then cover the other fourdesign methods and then return at the end to this,my main personal designmethod.Unlike the other design methods that I will be discussing it is hard toreduce this one to a set of rules or procedures,so it is best illustrated bysimply showing several design examples.The basic idea here is to file awayin your memory any interesting or unusual designs or principles of aberrationtheory that you run across and then,maybe many years later,to pullthat out and use it as the basis for some idea about a new type of design.Anequally important component of this design method is toconstantly look for and then question hidden assumptions that you or otherpeople have brought to the design problem you are working on.My first design example is exceedingly simple.When I was 15 years oldI bought the book"Mirrors,Lenses,and Prisms"by Southall.This was firstpublished in 1918 and has many things in it that have long sincedisappeared from optics texts.It is full of very complicated geometricof prism (n'>n)constructions and proofs,like this one.One simple picture stuck in my mindFio.78.-Experimental proofof minimum deviationand it is shown here.When a prism is being rotated to find the minimumdeviation angle in order to measure its index of refraction,the refracted raydeviation turns around at some point and reverses direction.It is hard todecide exactly when that happens since it is a stationary point with no firstderivative.This drawing shows how a parallel input ray reflected from theoutside base of the prism,a ray moving very fast as the prism is rotated,willbecome parallel with the refracted ray at the minimum deviation setting.This seemed very clever to me and I filed it away in my memory as anon-obvious way to use a simple prism.Later,when I was an optics studentat the Univeristy of Rochester,one of my professors was MVRK Murty.Hewas and is an extremely creative person.He devised some new ways to usesimple prisms and found some non-obvious ray paths through prisms.Thisfurther alerted me to the fact that simple prisms might still have someunexplored potentials.Now let us jump forward from back then to 1980.The Spanish surrealist artist Salvador Dali had recently mastered the art of