willis@rand.org (Willis H. Ware) (04/22/89)
Perhaps some of you can clarify with certainty a question that I have asked many telecommunicators and gotten unsure and/or waffling answers. It concerns bits and baud rates, and units. In careless usage, I notice (especially among computer types) the use of baud and bits per second synonymously - which I believe to be a corruption of the units. I'm familiar with the old mechanical TWX machines and their mechanical commutators, so I know that they transmitted a start pulse, 5 data pulses (a Baudot character) and a stop pulse per rev of the commutator and thus, per character transmitted. I also know where the name Baudot code comes from and the derivation of Baud from the proper name. My hunch is that Baud, in a rigorous sense, ought to be used only in connection with a character group of the telecom system in question and to designate the rate of transmission of such character groups, be they the 7-bit groups of the old TWX, or the 8-bit groups of byte-oriented systems. Thus the relationship between Baud and bps (I'll assert) will be variable and dependent upon the system under discussion, if the term is used with rigor and precision. In elaborate systems that incorporate error-control and other things, the bps and (for example) the bytes per sec will not differ by 8 but some larger number dependent on various technical details of the transmission format, including extra bits added for error control. OR maybe the accessory (so to speak) bits are never counted? Maybe a Baudot character is really 5 bits - ignoring the start and stop. Maybe error checking bits are never counter? Can you confirm any of this or point me to a precise definition of Baud? Of is this one of those things that the technical field is generally careless about? Incidentally, I occasionally run across someone that would like to use "elements per second" but it's a careless synonym for bits per second. On a related topic .. It seems to me that the "word" of Words Per Minute has always taken to be 5 characters which would make (say) 65 WPM equate to 325 characters per minute. If my argument above for meaning of Baud is accurate, then dividing by 7 (1+5+1) would yield a baud rate of about 46. OR maybe the divisor is 5 in which case the baud rate would be 65. I will appreciate any comments that you care to make. Thnx for your thoughts. Willis H. Ware RAND Corp Santa Monica, CA 90406 willis @ rand.org