braner@batcomputer.tn.cornell.edu (braner) (12/18/86)
[] Another challenge: A really fast "autocopy" program (to copy a list of stuff off a disk, upon boot-up, to a RAMdisk) could be written using the following idea: Put all the files to be copied into a pretty-much consecutive area on the disk. This can be done simply by writing them to the disk consecutively. It would also be helpful if they are in a special folder of their own. The program would read the directory of that folder, and the disk's FAT. It would then figure out from those the range of consecutive tracks that should be read to have all the necessary data in RAM, and read those AS ONE BIG BLOCK into a RAM area OUTSIDE THE RAMDISK. Later, the files would be extracted from this block and copied to the RAM disk. The hard part is making it independent of the specific RAMdisk program used, but for starters it would be great to have a version that works with the reset-immune RAMdisk (who would want to use any other?). (BTW: John Harris' RAMdisk and "eternal" are basically the same, as far as I can tell - they differ only in the way the requested RAMdisk size is sensed: keyboard vs. a data file). Philosophical note, partially due to Hal Hardenbergh of Digital Acoustics: Note that the ideas in this posting and my preceding one are rather obvious once you break out of the old mindset of small RAM, large disk. The day the first machine came out with more RAM space than disk space will eventually be remembered as a major event in the history of SOFTWARE! Why is it that two drives used to be absolutely necessary for serious use of a micro, and now many people are happy with a 1040ST with its one drive? You got it. Still wonder why I don't like the Mac? - Moshe Braner