I have no problem with people following such an approach but for most of us in the scientific community it doesn't work.
Consider two examples: mathematical software and graphics/GPU capabilities. Much of the mathematical software that is still crucial to many of us was developed back in the 50's and over the years has involved thousands of person-years of development. If that won't run on the latest hardware it is disastrous for most application users who cannot and should not have to go back and "fix" the libraries. The development of graphics hardware and software was originally driven by the scientific community. When graphics hardware moved to the chip level, games dominated because not only did the hardware become inexpensive but the game players were buying the software. The people represented on this list have gone from being major buyers of hardware and software to a very minor part of the revenue. If you look at the capabilities of GPUs, they are determined by what game players want. So, for example, those of us in the visualization community have to figure out how to use what we is available rather than having significant input into what will be available. If you want some further evidence, take a look at the membership of Kronos committees. The research and education communities are almost totally unrepresented on any of them and they are the ones that are setting the standards that will determine the next generation of hardware and software. Ed __________ Ed Angel Founding Director, Art, Research, Technology and Science Laboratory (ARTS Lab) Professor Emeritus of Computer Science, University of New Mexico 1017 Sierra Pinon Santa Fe, NM 87501 505-984-0136 (home) an...@cs.unm.edu 505-453-4944 (cell) http://www.cs.unm.edu/~angel On Feb 8, 2013, at 2:19 PM, glen e. p. ropella wrote: > Edward Angel wrote at 02/08/2013 12:57 PM: >> For many years >> graphics and mathematical software was driven by the scientific >> community which valued stability and backward compatibility. When the >> market became dominated by game players who are willing to replace their >> entire systems every year, the business changed dramatically, not only >> in terms of the software but also in terms of the hardware. > > Arg! You piqued me again. ;-) I know lots of gamers who don't merely > value backward compatibility, they go to ungodly extremes to maintain > old systems/games, port old games to new systems, build emulators for > old games, etc. > > But my pique isn't to argue about whether gamers value backward > compatibility. It's a common thread I've been pushing with regard to > scientific modeling and simulation (M&S). > > For better or worse, I've taken the stance that science using M&S should > be handled in the same way other science is handled. If you want to > reproduce a result, you don't slice out the repeatability at some > logically impermeable layer of compatibility. Instead, you keep (or > reconstruct) the _machine_ that was used for the original research. To > the extent that's not reasonable, then you run your experiments with an > alternative machine and characterize how the variation introduced by the > machine percolates into the variation in the results of the experiments. > The pure (math, physics, first principles) perspective of coming up > with _precise_ analogs for various machine parts is bizarre ... fine as > a fetish/avocation but inappropriate as a vocation. > > Yes, I know this is antithetic to most compsci people ... portable code, > universal turing machines, IEEE designed floating point representations, > etc. But, to me, it makes the most sense. As (non-universal) computers > further burrow themselves deep into our ecology, the concept of a > logical abstraction layer makes less and less sense. > > -- > glen e. p. ropella, 971-255-2847, http://tempusdictum.com > > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
