But I had a nice break when I went to Infovis 2006, the symposium on information visualization. There was a bit too much math this year, starting from the keynote, which was Eades talking about graph layout algorithms. I still managed to get something thought-provoking from it. His criteria for algorithm evaluation was "effective, efficient, and elegant."
These are good principles for software design as well as algorithm design: a good piece of software should be effective at supporting the tasks it's designed for, be efficient in use and for use, and ideally is elegantly designed. Elegance, of course, implies more than "usability." Usability is a word that's got kind of an old school ugly lab study connotation these days; it's a word that doesn't say enough to capture current thinking about the value of delightful design, rather than just adequate design, in creating a differentiating user experience.
What's "elegant" in a proof or theorem, I asked of a friend who was a mathematician in his previous life. "Simple," was the first thing he said. But not just that -- it can be taught to a second year student, was one of Eades criteria (suggesting "learnability"). Yet also somehow "surprising." An elegant proof is a result with a twist you didn't see coming, but should have, adds an insight that makes it aesthetically pleasing.
At risk of triteness, I did look up elegance on dictionary.com after striking out in a Google search: "gracefully concise and simple; admirably succinct. Combining simplicity, power, and a certain ineffable grace of design." It adds:
The French aviator, adventurer, and author Antoine de Saint-Exup'ery, probably best known for his classic children's book "The Little Prince", was also an aircraft designer. He gave us perhaps the best definition of engineering elegance when he said "A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away."
E3 makes a good compound principle for evaluating design of all things, including software. I'd really like more elegance in my design.
1 comment :
Elegance is something that seems to have different meanings to different groups... something that is clearly in the eye of the beholder.
Consider Dirac's formulation of quantum mechanics. To a physicist it is pure elegance - breathtaking elegance. It is a bit vague and ambiguous and allows you to grasp the physics content of the theory. It allows you to manipulate the physics content and find new physics. But this beauty and clarity is diminished if you worry about some uninteresting mathematical technicalities.
A mathematician looks at the same theory and only sees those minor mathematical issues and see something impure - even ugly. Of course they are interested in proof and rigor than the underlying physics.
Post a Comment