| This may just be me picking on one small part of the entire post, but I have a serious technical problem with some of the author's claims, not least of which is his feelings that FPGAs will somehow revolutionize personal computing. Anyone who has programmed with an FPGA will understand the sort of tasks they're best for; hand-optimizing particular algorithms and processes into parallel pipelines and execution units. Running a general-purpose OS on a CPU built in software on top of an FPGA is just not feasible; the yeild of transistors to logic units is very low for FPGAs compared to conventional microprocessors; and there's matters of efficiency, layout, and cost-effectiveness to think of. There has been some success using an FPGA-type programmable gate array in combination with a standard CPU, but the CPU itself is usually a proprietary core made to coordinate and assist the FPGA. These sorts of FPGA co-processors are only ever used for very specialized aplications currently. They won't be replacing general-purpose dedicated CPUs anytime soon (if ever). FPGAs are perfectly suitable for any special-purpose functions they've been programmed for: but I think in terms of desktop and mobile computing, having a farm of parallel execution units in the form of a modern GPU will also yield acceptable results for any parallel consumer applications without the need for as much investment as with dedicated reprogrammable FPGA hardware. The other technical concern I have is mesh networks; we don't currently have the technology to create and maintain large-scale mesh networks, but even if we did, there is a point there the communication required just to maintain the network outpaces the node-to-node bandwidth, or even the global bandwidth available. You also have to deal with privacy concerns around data storage and transmission at nodes while it is en-route to a destination, and circle-of-trust issues, etc. These issues bring the pendulum in a full circle. I agree that the pendulum does seem to be in a state of change at the moment, and all that is keeping it from swinging back (and probably giving it more overall momentum in the progress) is government and business interests. I think the issues we face, and that we have always faced, are not so much technical as they are social and societal. By the time we solve the issues around the commons, ownership, and governance; the technical issues will seem trivial. |