3 ms·

http://arstechnica.com/uncategorized/2007/07/research-optimal-copyright-term-is-14-years/ http://arstechnica.com/uncategorized/2007/07/research-optima... With

http://arstechnica.com/uncategorized/2007/07/research-optimal-copyright-term-is-14-years/ http://arstechnica.com/uncategorized/2007/07/research-optima... With copyright's explicit purpose of "promoting the sciences and useful arts", one can derive the ideal length. It must be long enough that there is incentive to create, but not so long that future generations cannot use it as a platform. Based on this professor's calculations, 14 years is ideal. 28 years, then, should be perfectly sufficient.

anigbrowl 12y ago
Oh, dear - that substantially overstates the conclusions drawn in the paper itself: http://rufuspollock.org/papers/optimal_copyright.pdf http://rufuspollock.org/papers/optimal_copyright.pdf The author warns that estimation of the ideal term depends on a number of imponderable factors like 'cultural decay' which are necessarily estimated, as well as discount rate for economic returns on copyrighted works. After considering a range of possibilities for these parameters through the medium of a probability density function, he observes that the median of the distribution is about 15 years - in other words, there's a 50% probability of that being the optimal term. The mode is about 20 years, ie the most likely suggested term for any random sample of the input value ranges. Going on, 'From the underlying cumulative distribution function we can calculate percentiles and find the 95th percentile at just under 31 years, the 99th percentile at 39 years and the 99.9th percentile at just over 47 years. This would suggest, that at least under the parameters ranges used here, one can be extremely confident that copyright term should be 50 years or less – and it is highly like that term is under 30 years (95th percentile).' So while this paper is great find, I think you need to walk back your conclusion significantly. Also, we need to be very careful about the initial assumptions on which the theoretical foundation is based: 'C is a variable denoting production costs of ‘originals’ (authorised reproductions) and U a variable denoting the production cost of ‘copies’ (unauthorised reproductions).' The author is essentially modeling the case of works-for-hire commissioned by a publisher and put on the market more or less immediately. His utility function aims to optimize the production of copyrightable works while maximizing public welfare (in terms of access to the largest possibly number of works), but it's important to note that he's using production in the sense of distribution of an already-created work by a rightsholder. This doesn't address things like creation time which may or may not be affected by technology (whereas technology has a huge effect on the reproduction of costs), or delays between creation and publication (which as I've observed elsewhere in the thread can be substantial).

anigbrowl 12y ago

Oh, dear - that substantially overstates the conclusions drawn in the paper itself: http://rufuspollock.org/papers/optimal_copyright.pdf http://rufuspollock.org/papers/optimal_copyright.pdf The author warns that estimation of the ideal term depends on a number of imponderable factors like 'cultural decay' which are necessarily estimated, as well as discount rate for economic returns on copyrighted works. After considering a range of possibilities for these parameters through the medium of a probability density function, he observes that the median of the distribution is about 15 years - in other words, there's a 50% probability of that being the optimal term. The mode is about 20 years, ie the most likely suggested term for any random sample of the input value ranges. Going on, 'From the underlying cumulative distribution function we can calculate percentiles and find the 95th percentile at just under 31 years, the 99th percentile at 39 years and the 99.9th percentile at just over 47 years. This would suggest, that at least under the parameters ranges used here, one can be extremely confident that copyright term should be 50 years or less – and it is highly like that term is under 30 years (95th percentile).' So while this paper is great find, I think you need to walk back your conclusion significantly. Also, we need to be very careful about the initial assumptions on which the theoretical foundation is based: 'C is a variable denoting production costs of ‘originals’ (authorised reproductions) and U a variable denoting the production cost of ‘copies’ (unauthorised reproductions).' The author is essentially modeling the case of works-for-hire commissioned by a publisher and put on the market more or less immediately. His utility function aims to optimize the production of copyrightable works while maximizing public welfare (in terms of access to the largest possibly number of works), but it's important to note that he's using production in the sense of distribution of an already-created work by a rightsholder. This doesn't address things like creation time which may or may not be affected by technology (whereas technology has a huge effect on the reproduction of costs), or delays between creation and publication (which as I've observed elsewhere in the thread can be substantial).