Newsgroups: alt.humor.best-of-usenet Ya know, This might just explain the signal problems that I've been having on my television cable. : : What's the concensus on the AT&T connectors as opposed to coax? I've : : heard all the Toslink-bashing, but it seems that AT&T would be fairly : : serious about good fiber connections. : Quite obviously, coax is the *IDEAL* medium for digital transmission : because it provides essentially *perfect* rejection of any spurious : non-digital signals. : When the size of the coax cable is properly matched to the font of : the digital source, 1's flow (lengthwise) down the center conductor : while 0's pass (broadside) down the shield -- and NO other numbers : can get through. : Unfortunately, many low-cost A/D converters and CD transports save : money by using cheap, public-domain fonts in which the 0's are oval : rather than perfectly round. The resulting null-distortion causes : unnatural (and easily recognizable) compression of the sound stage. : This effect can be partially corrected by using specially designed : coax with a slightly oval cross-section. : WRT the transmission of 1's, there are two competing bodies of : opinion. Most purists tend to favor the minimalist approach of : using a sans serif font with a solid center conductor; however, : some recent studies suggest that there may be real benefits to : a twisted center conductor in combination with a slight serif : at the base of the 1's. Apparently, the serif allows the 1's : to engage the spiral conductor and impart a stabilizing spin -- : in much the same way that the ogive at the base of an artillery : shell engages the rifling in the gun barrel. : hope this helps, : Mark : -- : Moderators accept or reject articles based solely on the criteria posted : in the Frequently Asked Questions. Article content is the responsibility : of the submittor. Submit articles to ahbou-sub@acpub.duke.edu. To write : to the moderators, send mail to ahbou-mod@acpub.duke.edu. =========================================================================== More on digital interconnects I forwarded your mss on 1's and 0's in coax around here, and got this response from Matt (I think you met him when you had lunch here recently). Isaac, It seems that there is some confusion over this issue. The concept that coax is better for digital data transmission because the 1's and 0's line up with the physical medium is utter nonsense. While coax may have yielded slightly better performance than twisted-pair in the past, this is not the case with modern digital data transmission systems that employ data compression. As anyone in the industry knows, bits of data are organized into groups of eight called bytes (from the Indoeuropean word for eight, "bytae"). With data compression, these bytes are what are transmitted over the physical medium, in lieu of the actual bits that comprise them, yielding a compression ratio of anywhere between 8:1 (in the case of a "1" sent) to infinity (in the limiting case of all 0's sent). The question now is "Why eights?" Here's where the genius of twisted pair comes into the picture. The eights are sent broadwise (a.k.a "bit-parallel, byte-serial") down the twisted pair, each loop of the eight encompassing one wire in the pair (electromagnetic waves don't actually flow through wires -- they flow through the space around the wires); since the wires are twisted, this imparts a natural spin to the eights. Furthermore, the greater the number of turns per inch, the higher the angular velocity of the eights, leading to a higher data stability. Of course, with more turns per inch, the eights are travelling a longer path, so this higher data stability is at the cost of a longer latency between data transmission and reception. With a half-duplex connection, this latency really limits the effective bi-directional data transmission rate; this is why full-duplex connections (for asynchronous bi-directional data communications, where the data being sent one way is not dependent on the data being sent the other way) yield data transmission rates greater than twice that for half-duplex. There are a number of excellent references on this topic that I can forward to you if you want. Regards, Matt =========================================================================== DPM> If zeros get converted to ones by the Fitzgerald contraction then DPM> do eights get converted to infinities? RM> Actually, each byte gets converted to infinity each time the pair RM> of wires is at the same level, reverting to an eight when the wires RM> are stacked vertically.