Basement Chronicles

Heuristics

"He turns off the slides, and it doesn't look right."

My computer networks professor defines a protocol as a system by which we can convey meaning between two entities. That is, a system of rules that define message format, transport method (optionally), and actions to perform on the event of sending and receiving a message.

Protocols, in terms of computers, are necessary to facilitate communication between one device and the next. Every computer has a set of programs that are designed around these protocols, programmed specifically to send messages using certain formats or to take certain actions when things happen. Not only does your computer "speak" any number of protocols, but each of these protocols is related to others, either having protocols built on top of the functionality other protocols, or protocols that are "aware" of other protocols and being able to react as a result.

Protocols, in terms of everything human, are not especially different. Our protocols are predicated on the same systems of information theory and physics. We even exhibit the same functional patterns and tricks in the design of our communicational protocols. At the most basic level, the only difference that I can see between the design and implementation of computer and human protocols is that ours are somehow mutable, while a computer's is not.

One would think that having a good understanding of this concept would aid me in understanding, implementing, and acting upon these human protocols. It doesn't in the slightest.

I have an example. Allow me to go into ridiculous detail:

The context is a large lecture hall full of engineering students. The teacher is communicating in American English on the predefined subject of statistics. As an (optional) header to the class, he passes around a small object (from his workplace, where he has mentioned on a previous occasion in the same context that his occupation uses statistics).

The protocol is thus: The first person to be given the object will look with genuine/feigned interest, based on whether the student does/does not know what the object is, respectively. The object, in this case, is the message, meant to convey information about a) itself and b) its relevance or relation to a previous or current lecture. The message is required to travel from the first entity's possession to the next's, until every entity in this context has received the message. Certain constraints must be put on this transfer, however. The entity (student) will not look at the object for more than the length of time in the class divided by the number of students in the class. The student, when finished with the object, will then pass it to the person next to them that has not seen it.

I could go on about the finer points of the protocol, detailing what happens when the object gets around the room, or when it reaches the end of a row, or when the object mysteriously/erroneously ends up in your hands and everybody around you has already seen it. Did I mention that error handling is part of protocol design as well?

So this protocol is based on another protocol: handing an object from one person's hand to another person's hand. A seemingly simple task, but one not to be trifled with when trying to provide an adequate protocol. The problem brings up interesting questions as to how to implement it. For instance, does the protocol allow for one person's hand to touch the other person's hand during transfer? What happens if it does? How does one negotiate gravity, friction, momentum, and any number of other physical forces in order to allow for a successful transfer?

Ah, so we find that in order to solve the larger problem, we need to solve the smaller parts first. So we need more protocols. And protocols on which to base those. And so on, until we find that we no longer need to delve further into sub-problems in order to solve the super-problems.

But that can't be it, can it? No, we can't just work vertically, moving up and down between different levels of abstraction. There are many more protocols in existence in this context, and not all are related to each other as being children or parents of the others. Some are running parallel, or more to put it a bit more visually, horizontally with the other protocols we've described. So instead of a neat chain of protocols to work with, we have a complex web of interrelated protocols that interact, affect, and sometimes conflict with each other.

And now we discover that this is quickly becoming a mess. The object-passing protocol, in addition to being obscenely complicated, is meant to be executed alongside the interpretation of the American English language broadcast interpretation protocol. Thus, due to limited attention span and multiprocessing ability of the human mind, the object-passing protocol takes a backseat, and eventually breaks down into complete chaos.

People drop the thing on the table, making a metal-on-particle-board clang every two or three people. Some people can't decide whether to drop the object into the other person's hand (gravity utilization protocol), or to hand it to them in some complicated "synchronous doubly-clasping object-transference protocol" (real name unknown), or to simply roll the object between them and hope they are aware enough of the situation to catch it (rolling protocol).

A few other patterns arise. Some people, for instance, make eye contact with the person they are passing the object to before they initiate one of the object transference protocols. Some people laugh at the object. Some people make thoughtful noises towards the object, usually when turning it over and over in their hands before passing it along. Some people initiate a short conversation (or pick up a previous one), referring possibly to the object they are handling.

Some people, like me, sit quietly, ignoring the lecture, and observe the process. I wonder what action I am going to take when the time comes for me to enact upon any number of predefined protocols. I realize, then, that my actions are determined partially by the person who will hand me the object. And those persons actions are defined by the person previous to them, and so on, with each interaction being a product of itself and every previous action.

I realize, then, that a girl from my lab is sitting next to me, and she will be the one handing me the object. I also realize that I probably smell pretty bad, considering I bike for half an hour to get to this class, and it is 105 degrees Fahrenheit outside. In fact, I'm covered in sweat, my hair is sticking up, and I'm dripping. The front of my shirt looks like the scatter plot on the front screen, with the mean at the center of my chest, and 68% of the dots are within 4 inches of the mean.

Damnit, this is getting more complicated. Now, if anybody else is thinking like I am, it is possible that those person's actions are determined not only by the person that will be handing you the object, but also by your preconceptions about the person you're handing the object off to. The permutations are giving me a headache.

She gets the object. She spins it once, makes no sound, and then hands it off to me. She tries to make eye contact.

I maintain eye contact with the object, spin it around, make a quiet "hmm" sound, and pass it along, and continue making recursive drawings on my notebook.

I realize then that there is no protocol. Either that, or that the only way to exhaustively define such a protocol would be so difficult it might as well be impossible. There is no system of rules and definitions. There is nothing for me to understand about it, nothing I can study and dissect and master and memorize. But how can we accomplish anything, even something as simple as passing an object around the room, without understanding it? And how can we understand it without understanding how we understand it? And so on, and so on.

So, I guess that our whatever algorithms we use to process our protocols are "heuristics" Approximations at best, but not always accurate or consistent.

So then what historical protocol analysis algorithm is telling me now that I should have made eye contact, smiled, and said thank you when the girl passed me whatever stupid piece of germ-covered metal?

Maybe I should stop trying to systematize everything. It'll eventually drive me nowhere but crazy.