Human Scale Part I - Planned Obsolesence
Having just finished Kirkpatrick Sale's Human Scale, I thought I might post my impressions. Although I'm sympathetic to the overall thrust of the argument, I think there are a few problems. Perhaps they are just with this book and the fact that some of it is dated. In any case, they don't seem to be entirely fatal flaws, but should be noted nonetheless.
Planned Obsolescence
This is a minor thread that runs throughout the book, usually in the background. It is one of Sale's fundamental assumptions about state capitalism. In one case, he describes the Century light bulb, about which I have already written. In other cases, he writes about the planned obsolesence of appliances such as refrigerators, with the implication that they are built that way to force you to buy more often. Let's take the lightbulb first.
The Century bulb is still running today in a firestation in Livermore, CA (the webcam is here). The bulb is never turned off, and for good reason: most (99%) lightbulbs don't burn out in normal operation, they pop during the first few milliseconds of being turned on. The current inrush creates a heat differential and therefore rapid and uneven expansion of the filament, causing it to pull itself apart. Thus, if you never turn it off, you never have to risk turning it on cold. The Century bulb itself is somewhat special, too. At 4 W, you can barely see it even when the other lights are off. This is so because the filament is huge.
You could obviously make bulbs last longer through a variety of techniques: thicker envelope, thicker filament, better evacuation, electronics to control the inrush current. The question is whether the changes are worth the extra cost. One $0.50 bulb every 2 years is no more economical than a $1 bulb every four years. Indeed, as the price of most goods declines in terms of the amount of work required to earn the money to buy them, even that simple comparison works out in favor of the shorter bulb over time.
The other claim - that appliances could be made to last longer but are intentionally not - is based on two mistakes. The first is based on a misunderstanding of statistical quality control (SQC). We can, after analyzing lots of appliances over time, figure out that an appliance will fail in a predictable manner. The failure probability looks like a bell curve. From that, we can say that Refrigerator X will last on average Y years. From this, people will infer that the refrigerator was designed to fail in Y years. In a sense, it was, since the refrigerator was designed within certain constraints: existing technology, cost points, market demand, competitive expectations, cost of inputs including capital and materials, etc. The end result of those design choices is a refrigerator that lasts, on average, Y years. But the direction of causality is from the design to the durability, not from a selected goal of durability to the design. This is a misapplication of statistics, and is usually committed either out of malice or ignorance. I'll assume Sale does so out of the latter.
The other mistake is the idea that people should design 50 year refrigerators (or whatever). Keep in mind that you can, right now, buy outstanding appliances from companies like Viking. They are very expensive. At the same time, keep in mind the fact that technology is changing and that the rate of change is increasing. Given both of those, why would you want to pay extra for something that will be overtaken by scientific and engineering - not design - obsolescence within a few years? The examples are mind-boggling: a car radio of a few years ago does not have as good reception, disc capability, or perhaps even cassette playback capability; the incandescent lightbulb has been overtaken by the CFL and is about to be overtaken by the LED; a state-of-the-art computer from 1990 won't even begin to approach the capability of a modern computer for most of the modern applications (such as the internet, USB, etc.); the most economical and reliable car from 1975 won't even touch the most economical and reliable modern car for either of those measures or for safety (remember when airbags were only available on high-end Mercedes?). So why would anyone pay a premium for that which they could have in the future at a deep discount? Now, Sale would probably claim that all of those are examples of products designed to function within the state capitalist system, but what about that panacea of solar panels? The efficiency was pitifully low in 1980 (when Human Scale was written) - were they planned to be obsolete? Isn't it good that the efficiency of commercial units is as high as 20% today? They have lab experiments returning 30% efficiency - don't you want that in the future?
Planned Obsolescence
This is a minor thread that runs throughout the book, usually in the background. It is one of Sale's fundamental assumptions about state capitalism. In one case, he describes the Century light bulb, about which I have already written. In other cases, he writes about the planned obsolesence of appliances such as refrigerators, with the implication that they are built that way to force you to buy more often. Let's take the lightbulb first.
The Century bulb is still running today in a firestation in Livermore, CA (the webcam is here). The bulb is never turned off, and for good reason: most (99%) lightbulbs don't burn out in normal operation, they pop during the first few milliseconds of being turned on. The current inrush creates a heat differential and therefore rapid and uneven expansion of the filament, causing it to pull itself apart. Thus, if you never turn it off, you never have to risk turning it on cold. The Century bulb itself is somewhat special, too. At 4 W, you can barely see it even when the other lights are off. This is so because the filament is huge.
You could obviously make bulbs last longer through a variety of techniques: thicker envelope, thicker filament, better evacuation, electronics to control the inrush current. The question is whether the changes are worth the extra cost. One $0.50 bulb every 2 years is no more economical than a $1 bulb every four years. Indeed, as the price of most goods declines in terms of the amount of work required to earn the money to buy them, even that simple comparison works out in favor of the shorter bulb over time.
The other claim - that appliances could be made to last longer but are intentionally not - is based on two mistakes. The first is based on a misunderstanding of statistical quality control (SQC). We can, after analyzing lots of appliances over time, figure out that an appliance will fail in a predictable manner. The failure probability looks like a bell curve. From that, we can say that Refrigerator X will last on average Y years. From this, people will infer that the refrigerator was designed to fail in Y years. In a sense, it was, since the refrigerator was designed within certain constraints: existing technology, cost points, market demand, competitive expectations, cost of inputs including capital and materials, etc. The end result of those design choices is a refrigerator that lasts, on average, Y years. But the direction of causality is from the design to the durability, not from a selected goal of durability to the design. This is a misapplication of statistics, and is usually committed either out of malice or ignorance. I'll assume Sale does so out of the latter.
The other mistake is the idea that people should design 50 year refrigerators (or whatever). Keep in mind that you can, right now, buy outstanding appliances from companies like Viking. They are very expensive. At the same time, keep in mind the fact that technology is changing and that the rate of change is increasing. Given both of those, why would you want to pay extra for something that will be overtaken by scientific and engineering - not design - obsolescence within a few years? The examples are mind-boggling: a car radio of a few years ago does not have as good reception, disc capability, or perhaps even cassette playback capability; the incandescent lightbulb has been overtaken by the CFL and is about to be overtaken by the LED; a state-of-the-art computer from 1990 won't even begin to approach the capability of a modern computer for most of the modern applications (such as the internet, USB, etc.); the most economical and reliable car from 1975 won't even touch the most economical and reliable modern car for either of those measures or for safety (remember when airbags were only available on high-end Mercedes?). So why would anyone pay a premium for that which they could have in the future at a deep discount? Now, Sale would probably claim that all of those are examples of products designed to function within the state capitalist system, but what about that panacea of solar panels? The efficiency was pitifully low in 1980 (when Human Scale was written) - were they planned to be obsolete? Isn't it good that the efficiency of commercial units is as high as 20% today? They have lab experiments returning 30% efficiency - don't you want that in the future?
Labels: book, decentralization
<< Home