I couldn't figure out how to reply to the main article so I am replying to a reply.
Slow transitioning from pure to applied is a common problem in all of science. An exception that might help prove the rule is the theory of wavelets. Early research in the field was done by pure mathematicians, signal processing engineers and others. All the researchers coordinated to get wavelets into common use. Prizes were awarded.
I suggest the following ways to speed up the transitioning process in any field of science:
1. Transitioning should be rewarded with promotions, prestige, prizes and grants.
2. Software should be well-written, well-documented and under an open source license. Equivalent things should be done for laboratory techniques, etc. Reinventing the wheel is expensive and slow.
3. Both the pure and applied literature must follow the academic tradition of openness. It is impossible to coordinate with someone whose work is secret.
I open my copy of Digital Image Processing by Pratt and I find on page 4:
Y(x, y, t) = integral from 0 to inf C(x,y,t,l)V sub S(l) dl
On page 6 is the Dirac delta quickly followed by Fourier transforms and a little linear systems theory.
Looking throught the entire book I see a lot of mathematical analysis, linear algrebra, linear systems theory, probability theory,... My copy of the book was published in 1978! Nowadays thing are worse.
It is easy to understand why a computer science professor once told me that most of his graduate students were poorly prepared in math.
Slashdot Mirror
I couldn't figure out how to reply to the main article so I am replying to a reply.
Slow transitioning from pure to applied is a common problem in all of science. An exception that might help prove the rule is the theory of wavelets. Early research in the field was done by pure mathematicians, signal processing engineers and others. All the researchers coordinated to get wavelets into common use. Prizes were awarded.
I suggest the following ways to speed up the transitioning process in any field of science:
1. Transitioning should be rewarded with promotions, prestige, prizes and grants.
2. Software should be well-written, well-documented and under an open source license. Equivalent things should be done for laboratory techniques, etc. Reinventing the wheel is expensive and slow.
3. Both the pure and applied literature must follow the academic tradition of openness. It is impossible to coordinate with someone whose work is secret.
Looking throught the entire book I see a lot of mathematical analysis, linear algrebra, linear systems theory, probability theory, ... My copy of the book was published in 1978! Nowadays thing are worse.
It is easy to understand why a computer science professor once told me that most of his graduate students were poorly prepared in math.
Homework: Study the most recent MPEG specs.