On Mon, 1 Jul 2019 at 15:06, Charles Mills <charl...@mcn.org> wrote: > > You're right, of course. Not to start a religious war, but even on a > big-endian machine, it seems to me to make sense to number > the bits from LSB to MSB. Bit n then represents 2^n -- in an 8-, 16-, 32-, > 64- or 128-bit integer. What could be more logical?
I think you're effectively proposing some sort of mixed-endian scheme. With big-endian addressing, no matter the size of the chunks of data, the numbering starts at 0 and increases "left to right". (Obviously you can write these things on paper any way you like - top to bottom or bottom to top, etc., but it's *consistent*.) So the first bit in main storage is numbered 0, the next higher one is 1, and so on. Byte 0 is the first, 1 the next, etc. Same for halfwords, fullwords, doublewords, quad words, pages, and any other grouping you like. Once you introduce bit numbering that goes the other way, is there a consistent reason not to do the same for e.g. page numbers? And page sizes *have* increased over the years, so it's not a hollow argument. > Not going to happen of course. Yeah. It's an annoyance, but better than the alternative... Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
