> > Bill Atkinson, of Apple QuickDraw fame, called > Genuine Fractals "snake oil". ie, bogus.
Well, it's certainly over-hyped. I've seen several people say "it's OK as long as you don't enlarge the image too much". Well, the same is true of regular bicubic extrapolation followed by a light application of unsharp masking. It does trade off some undesirable effects of digital content creation for artifacts that are less visible to the human eye, but don't be fooled into thinking that it isn't inventing the extra data just like any other digital extrapolation. Personally I wouldn't touch the software with a ten foot pole, but that's nothing to do with the way it performs. There's a lot of ill-will towards this package in the computer graphics community (especially amongst us old timers), dating from the initial announcement of the technique. Michael Barnsley used his allocated timeslot at SIGGRAPH to give a marketing puff presentation totally free of technical content to a room full of geeks interested in looking under the hood. Not only that; the promised forthcoming papers describing the algorithms in detail somehow never showed up in subsequent years. I was sitting amongst a group of computer graphics luminaries immediately after that paper, and Barnsley was excoriated in no uncertain terms. Not that this bothers him, I'm sure; Iterated Function Systems probably made a goodly sum of money.
