The Computer of Theseus

Having been buoyed by my success in making a Windows Home Server, I embarked upon a project to fix my mother-in-law's computer. Before I proceed too much farther, I ought to say that my mother-in-law had never asked me to do this. Indeed, quite the reverse. The computer was "working fine" she said. Clearly, she was just in denial. The computer was several years old. How could it possibly still be working in a satisfactory manner?

So I bullied her into allowing me to fix it. All I needed to do was replace the motherboard--and the CPU. I already had both of these left over from another project. It wouldn't cost a thing.

Then I opened her case. It was a genuine IBM. It was an old computer of mine. At some point, I'd treated myself to an IBM just for the bragging rights. This was before IBM got out of the PC market entirely. The case was nice. So nice it took me 15 minutes to figure out how to open it. This is one thing I've discovered: no two computer cases open in exactly the same way. The engineers apparently take delight in finding clever ways to keep end-users from getting the machine open. You have to proceed with the idea that opening will be easy—once you figure out how to do it.

Once the case was open, I began to feel fear. In making my home server, I'd worked with my wife's Dell which was new (a Celeron/Pentium 4) relative to this IBM Pentium II. My wife's Dell had a smaller motherboard. It was a micro-ATX (mATX) as opposed to the full-blown ATX sitting in the IBM. The IBM had a little fan in a special slot and pointed directly at the IBM's CPU. There were scads of wires running everywhere. One thing that I've learned in this new hobby is not to be daunted by the profusion of wires. The inside of a computer are composed of a relatively small number of pieces and these can be dealt with one at a time.

I began taking the pieces out a few at a time. I removed the PCI cards, the CPU fan, and the memory. Then I removed the motherboard. I put in the new motherboard and its new memory. The PCI cards I'd removed had been for a modem and an Ethernet card. My mother-in-law is on DSL, so she doesn't need the modem, and the new motherboard has an Ethernet card on it, so neither of those cards was needed.

The next step was attaching the wires to the motherboard. These come in three types: power cables from the power supply, hard drive cables, and front panel wires.

It is here that I want to make a digression in order to transmit a discovery of mine. This is in regard to the front panel wires. These wires have a number of separate functions. They connect to the hard drive LED on the front panel of the computer to tell you if the hard drive is working; they connect to the reset switch what allows you to reset the computer; and most importantly they hook you to the on-switch. I'd discovered while turning my wife's computer into a home server that the on-switch (the power switch) is the only one of these you really have to have. My discovery is there is no universally accepted way for these to be configured. On the newer motherboards I'd looked at, there was a single front-panel block on the motherboard. On this ancient IBM, the plugs were scattered from hell to breakfast. Furthermore, one of these wires which was a three-prong on the old motherboard had mutated to a two-prong on the new one. This change caused me to spend $21 on replacement front panel wires which were, ultimately, not needed.

Having bought the new wires and hooked the new motherboard up, I turned it on. Nothing happened. The fan didn't come on; the hard drive didn't whir to life; the setup screen did not popup on the monitor.

Rather than admit defeat to my mother-in-law, I read the manual to my motherboard. On it, I found a place giving an explicit warning that the motherboard wouldn't boot with a standard ATX power supply. It required an ATX12v power supply. I didn't know what kind of power supply the old IBM had in it, but I had hope that it wasn't the ATX12v required. The answer at this point was to purchase such a power supply, but this raised a question as well: would the new power supply fit the old box? I chose to cut to the chase and get a new case at the same time.

I did this. When the new case and power supply arrived, I installed the power supply in the case and then went over to my mother-in-law's house. I then moved the motherboard, memory, and hard drive over to the new case. At that point, the only piece of the original computer left was the hard drive.

With everything installed, I turned the machine on. The fan started spinning, the hard drive started whirring, and the setup screen popped to life on the monitor. My heart soared like a hawk.

Then there was a series of beeps and it turned itself off. I turned it on and watched closely. It did the same thing.

At this point, I replace the old IDE hard drive with a new SATA hard drive, thinking the old hard drive might be too old for the motherboard. (The hard drive was jumpered in a way that was odd to me.) I turned it on again; same thing.

I did, however, note that there was a message on the screen stating that the previous time the computer had been shut down it was because of a "thermal event." I've ordered a heat-sink and a fan for the CPU.

I am, after all of this, in a philosophical mood. This reminds me of the philosophical question on the ship of Theseus. Theseus has a ship. As the planks rotted, they were replaced. After all of the planks had been replaced, is this still the same ship?