On Wed, 6 Dec 1995, Jaco van Iterson wrote: > As a matter of fact a char consists of exactly 8 bits (a byte), numbers are > stored in a char by setting this bits on or off. Each bit has a value and > by adding this values together you get the number the char represents. > The values of the bits are: 1, 2, 4, 8, 16, 32, 64 and 128 (or -128 for signed). > Other ways to represent this numbers are: 2^0, 2^1, .... 2^7 (or -2^7). > and 1<<0, 1<<1, .... 1<<7. I prefer this last one (same for signed and unsigned) *argh* * I know how the bit system works! * the RBB's were intended to be an *emulator* of bits; if you're going to remain confused, I suppose I could call it CFS(Compacted Flag System), but I can call it whatever you need as long as you are aware that I *know* how many bits are in a byte, how many bytes in a word, et etc on non-supercomputer machines. I also know how bitfield operations work. This would be an _em-u-la-tor_, which means that its operations would _mimic_ that of bitvectors and flags, and would _also_ try to keep said emulation in as short of an amount of memory as possible. I am aware also of the tradeoffs between efficiency and speed and memory. This is an age-old debate. I'm not a new coder. Please recognise this and speak to me on the basis of that knowledge. Whether or not this system would impact speed or efficiency, I am choosing *at this moment* to center in on what would least impact memory. In such a case, making an algorithm to not have 8 flags take up 1 byte would be very nice. Now, are there _any_ workable ideas for this, or should I just ditch it based on the grounds that it's not possible? Work with this on the grounds that memory _is_ a factor, and speed/efficiency are less in the priorities. -- Kenneth G. Cavness | http://ccwf.cc.utexas.edu/~cavness Associate Editor | "That which is possible is not always MIDS, TIC | probable." -- Isaac Asimov 1-512-451-7602 | "What about the Tuna?" -- Unknown
This archive was generated by hypermail 2b30 : 12/07/00 PST