John S. Ward wrote: >I suspect you may be a bit out of touch on the Byte/bit thing. > >First off, the use of non-integral bits is quite correct, and certainly >possible.
We were talking about bit as the term is used in the computer field -- meaning binary digit -- having two states, on and off, usually represented as 0 and 1. A second-level storage element (e.g., a byte) must have an integral number of bits. The use of the same word in a purely mathematical context is irrelevant to this particular discussion. >This sort of thing is commonplace in mathematics. You >can raise a number to a non-integral power. Ever wonder how do you multiply something >times itself, e.g., 2.6 times? You can take the factorial of a >non-integer. >And you can have a non-integral number of bits. This situation >often arises >because computers are binary, but the real world isn't. The fact is that, on a computer, all numbers are represented within fields in a memory structure that has an integral number of bits, both overall and within arbitrary storage elements, such as the byte, and within whatever length field is being used. >Secondly, in English I've seen the symbol "b" used very >consistently for bit >and "B" for byte. The main exception seems to be non-technical people who >don't completely understand the difference between a bit and a byte. Yes, b is often used for bit, although in an incorrect manner, as in bps, kbps, Kbps, Mbps, etc. On the other hand, standards organizations, such as ISO and ITU, use bit, as in bit/s, kbit/s, Mbit/s, Gbit/s, etc. (Although I don't have a current version, I have the entire set of ITU-T [formerly CCITT] I-series, V-series and X-series Recommendations.) >I do appreciate your pointing out that this isn't an international >standard. And with good reason. As Robert Bushnell pointed out, B is the SI symbol for the acceptable non-SI dimensionless unit, bel, and, as I pointed out, b is the symbol for the acceptable non-SI unit, barn. Even though bits, bytes, etc. are non-SI, SI notation provides an excellent and consistent framework for expressing them. Bill Potts, FBCS, CMS Roseville, CA http://metric1.org [SI Navigator]
