Logic:Using RAM as swap
From Tractatus
I came across this fascinating article today,
Which beautifully illustrates some of the problems which occur when you reason from a false premise. It is not that the author has a low IQ. Notice how the author continually mentions that the hypotheses will need to be tested: " No more speculation and a priori knowledge of what might happen, but observation of what does happen." The author carefully takes a defensive position, and discusses different attacks on the premise ('here is how you use RAM for swap space').
Of course, the problem is that the poor soul has decided to take an experimental approach to determining how a designed apparatus works, without considering what the design actually is.
For those of you not computer literate, 'swap space' is space allocated on some large capacity device (Usually your hard drive), to prevent you from running out of memory in which to run programs. The idea is to allow applications to consume -as a whole- more memory than is physically available on the system. In addition, memory not recently used, can be swapped to disk, permitting better performance.
Essentially, your system is pretending that it has a total amount of RAM equal to (Amount of RAM) + (Amount of Swap Space). It pays attention that you can only have (Amount of RAM) actually resident at a time, and tries to minimize the number of I/Os to swap.
Using RAM for swap is... a very odd idea. You would rather use the RAM as more RAM, so that you did not need to swap at all. In fact, at that point, you might as well turn swap off entirely.
What is logically interesting about this argument is that it is similar in form to a reductio ad absurdum. First, we assume a hypothesis, and then argue about it. We use it to prove an absurd statement, and conclude that the hypothesis must be false. This process can be repeated ad infinitum, which is to say, a number of times equal to the number of provably false conjectures.
The author assumes that swapping is causing a performance problem, and that the individual should spend $10,000 on a solid state device to use for swap space. He then proceeds to argue that the money might be better spent buying RAM, and using the RAM to create a virtual disk, to use as swap. At this point, the logical conclusion is that the hypothesis is false, and that the individual should not purchase the solid state device.
Now, of course, the very idea of using RAM for swap is absurd. But, because the author used it to say something meaningful (Buy more RAM instead of a solid state device), the author incorrectly falls into the trap he set for the hypothesis. Now he starts to believe that the argument itself is valuable. He talks about how this setup should be tested. What he needed to do was think about what swap is, namely, a substitute for RAM.
All this goes to illustrate something about the form of the argument. When talking with people trained in mathematics, this method of argument is very natural. They will state, as if true, something they believe to be false. Then they will argue at length, as though they believed it. However, when they discover that it allows them to prove something false, they discard the hypothesis without a second thought.
I have noticed that sometimes those not as familiar with the style of argument will believe that the mathematician believes the hypothesis to be true. This can often result in anger, that such a statement would be believed, making further discussion impossible. Also, intermediate statements, which are also false because they rely on the same hypothesis may inadvertantly be believed.
It just goes to show that you need to be careful when making a counter-argument not to accidently believe it.
--Iain 15:20, 7 Sep 2005 (EDT)