Waste is a relative concept: "one man's trash is another man's treasure." In the early-90s, I remember frequenting a computer shop outside Columbus, Ohio that sold reused, recycled or "spare" computer parts. It was called Computer Success and it was the first technology thrift-store I'd ever seen. I still haven't come across another place like it.
These guys had buckets of computer parts, many long obsolete (even by mid-90s standards). Imagine those big bins at flea markets filled to the brim with 14.4 kilobaud modems and generic Trident VGA video adapters. Piles of yellowing, cigarette stained monitors that had spent most of the 80s in grungy office buildings. Cases, adapters, cables, processors, if it was made of silicon, they had it.
Most of this stuff was easily deemed "junk." No one went there to outfit a new company with cheap computers, you went there when you needed to patch something up. It would now probably be more profitable to recycle these components for minerals, rather than resell as used equipment. But in 1995, there wasn't much of a green movement and recycling computer equipment was rare.
Today, computer and gadget technology abounds. Millions of Americans have cell phones, almost as many have a computer, and a few have more than one. As technology becomes more prevalent, the waste piles up. A store like Computer Success in 2008 could easily pull most of their inventory out of the trash.
And it's not as though we're throwing away 50MHz 486 computers either. In 1992, the average 486 performed 0.818 million instructions per second per MHz. Five years ago this number was close to 3.0 MIPS per MHz. Not only has MIPS per MHz tripled, but MHz has increased by a factor of 50:
| Processor | Year | MIPS per MHz | MHz | Total IPS |
|---|---|---|---|---|
| Intel 486 | 1992 | 0.818 | 66MHz | 54 Million |
| Athlon XP | 2002 | 2.967 | 2.0 GHz | 5.9 Billion |
| XBox 360 | 2005 | 2.0 | 3.2 GHz | 9.6 Billion |
| Intel Core2 | 2006 | 9.242 | 2.93 GHz | 27.079 Billion |
When you look at those numbers, you realize the recent increase in computing power is staggering. In 1992, many people were able to live highly productive lives being able to execute no more than 50 million instructions per second. By 2002, in theory, one could be 11,000% more productive by executing almost 6 billion instructions per second.
My question is what happened to all the old equipment? A 2002 vintage machine would have been in use for 6 years by now. It has likely been upgraded or perhaps sits idly in basements. It couldn't run Vista or Leopard, but it's certainly still workable for someone in some part of the world.
The people at One-Laptop Per Child are utilizing this astounding increase in computing power. By providing a basic processor (400 MHz) and simple, free software, they've created a low-cost, high-function platform for any one. Given the statistics and anecdotal evidence about the volume of technology waste, there could be a lot more of this techno-charity going on.
Along these lines, my friend Carson, who is currently serving the Peace Corps in Gambia, setup a computer lab in his village last May (pictured below).
He writes, "Now that Kerewan has electricity in the mornings, Harinder and I have set up a computer lab in the school library with some old Pentium IIs that were lying around town collecting dust. We currently have 6 up and running with plans for a few more. The students, most of whom have never set hands on a keyboard, are begging for time and instruction. Between classes I've been informally teaching as many as possible, but it gets a little crazy with 50 kids trying to push onto 6 computers. On Monday we're starting evening classes for teachers, who are also first-timers, and I'm lobbying for formal student classes during the next academic year. This makes the expansion of the lab a top priority."
A great metric to have, in addition to things like per capita income, would be per capita computing power. Average instructions per second per household, divided by the number of households, perhaps. This could be an interesting way to re-frame the poverty debate. Most Americans realize there are poor nations in the world, but if you go around asking them for money they tend not to listen as well. However, if you frame the conversation in terms of computing power, whose disparity is likely to be significantly greater than income, you may get a more interested public.
There are still huge hurdles in techno-charity. I've only talked about obtaining hardware. Software is a little better with the amount of free and open-source applications available. But then there are additional technical problems, including electricity and connectivity. And that's all before you even get to the biggest issue: instruction, education and training.