Since Rali still seems to be gone, and I want to keep her blog going, how about a little cross-disciplinary thinking?
Most people know by know that computers think in zeros and ones. Every letter you type on a keyboard, every video you look at on YouTube, every sound your computer produces is represented internally by a binary number. For example, the letter “A” would be stored in computer memory as 01000001. Of course, the computer doesn’t actually store a number, but a representation of a number. These representations can take a variety of forms, such as a high or low voltage in a chip (High representing 1, low 0.) When signals are sent across buses and between computers, a pulse of voltage at a certain level represents a 1, whereas a lack of voltage represents a zero.
Below is a waveform I captured with an oscilloscope off of a T1 line in my home lab. A T1 is a type of digital circuit provided by the telephone company, which is commonly used to connect computer networks over long distances. I captured two waveforms simultaneously, one off the transmit side and one off the receive side. Look first at the bottom graph. You’ll see a series of pulses, some going up, some going down, with an occasional flat line between them. This would be interpreted as 1111101111110… Each pulse represents a one, regardless of whether it points up or down. This alternation of the pulses between negative and positive is a deliberate design which creates a zero net DC voltage on the line. The important thing to notice is that the pulses always alternate, with positive following negative and vice versa.
Now look at the top chart. We see a pattern that looks like 1101100011011001… But we clearly see two negative pulses, followed shortly after by two positive pulses. This is a violation of the rule I just described. What gives?
It turns out that T1 lines don’t like to have long strings of zeros. It throws off the timing of the line. So whenever a long string of zeros shows up at the T1 equipment, it throws in a special pattern of 00011011 to remove the zeros, but then forces two bipolar violations to signal the receiving equipment that what it really wanted to send was 00000000 and not 00011011. If it had really intended to say 00011011, the signal would have normal, alternating polarity. We call this Binary 8 Zero Substitution, or B8ZS.
So what does this have to do with anything? Well, the devil is in the details, whether it’s computers speaking or humans. It shows how two things can look remarkably similar, and yet be completely different. For Greek students, it should emphasize the importance of paying attention to diacritical markings! Just like our 00011011 looks like 00011011, but is in fact 00000000, ἤ may look like ἡ or ᾖ, but they are not the same thing. ποίει looks almost identical to ποιεῖ, but the accents are different, and so are the meanings. The Greeks may not have needed diacritical marks, but they are your best friend, so keep your eye on the less obvious details if you want to read Greek. Learn the diacritical marks as a part of the spelling of the word, and pay attention to those iota subscripts, which, although not technically diacritical marks, are treated the same way by many students, and ignored.