Interchangeable Parts
an essay about technology
This essay is a time machine.
I wrote it in 1998 when the World Wide Web was 3 years old.
I was thirty three years old, just married, and lived with my wife in an outer Noe Valley flat downstairs from a perfectly eccentric landlady.
I owned a small software business. As part of a content-driven marketing strategy to grow my business I wrote a collection of articles to post on my website. But then I got an offer to sell my business and work on a streaming music startup with Toni Schneider and I decided to do that instead.
So I stored the articles in a folder of stuff that I then religiously shifted along with each evolution of my data backup strategy from Zip drives to external hard drives to an early S3 reseller named JungleDisk. It sat there until 2015 when I decided to move the folder once more to Google Drive, and in the process read some of the old files.
When I read this one brought me right back to the dawn of the internet where I could say charmingly naive things like “…in the end Apple may not have as much affect on the computer industry as Henry Ford did on the auto industry… .”
It turns out it took longer than I expected but interchangeable parts for software, software ubiquity, what I was predicting in this essay, has happened. As I expected, as we stopped having to spend all our time just getting websites to work, websites went from the hottest thing into a background technology we take for granted to get things done.
Which, in turn, creates room in our minds to obsess about the next big thing — AI, self driving cars, clean energy—in the extremely human drama of looking for technological progress and ending up with technological change over and over again.
Eli Whitney invented the cotton gin, the device which made possible the profitable production of cotton fiber. You learned this in grade school, right? But did they also tell you, like they told me, that Eli’s real claim to fame, or anyway his most lasting contribution, was actually the concept of interchangeable parts?
As a kid this bothered me: the cotton gin was an invention which immediately and profoundly changed the American South; interchangeable parts was just a trick which meant Eli could make his gin a little cheaper. How could this obvious idea be more important than the complicated machine that people had been trying to invent for years? And anyway, I didn’t see how you could invent interchangeable parts. What other kinds of parts are there? Who would want a non-interchangeable part?
My confusion on this point continued until my teen years when I started working at a garage down the street. Charlie Graham had started this place only about 20 years after Henry Ford built his first Model T, and there was still plenty of evidence lying about of what auto repair had consisted of when cars were a novelty and Mill Valley was the boonies.
Back then auto parts, like autos, were scarce, expensive and of questionable quality. Rather than depend on the distributor in San Francisco for parts or advice, Charlie fixed cars by making replacement parts with the machine tools in his shop. Charlie was a great self-taught engineer and a warm, gregarious, friendly bear of a man who enjoyed making his customers happy. As a result he tinkered with their cars to the point that they often left his shop working better than when they left the factory (I know this because he told me stories and I saw him occasionally make improvements on cars when I worked there).
Fifty years later when getting grease under my fingernails was the coolest thing I could imagine the situation had changed completely. Car design had matured to the point where the machines delivered impressive performance and reliability. The parts distribution system was equally sophisticated; Mill Valley had three auto parts stores, and we could get pretty much any part for any car, even a car made in Italy, Germany or Japan, within a couple of hours. During the fifty years Charlie had been working, around him had grown up a world-wide integrated system that included every aspect of the design, manufacture, distribution, use, maintenance, and repair of automobiles.
This system hadn’t really been planned, though certainly some parts of it had such as the US National Highway System, but instead had evolved organically and incrementally as the result of millions of people adopting auto technology into their lives. Charlie and, for the last forty years, his brother Ed had been a small part of this organic development and when I started working at Graham’s Garage I added my energy to the mix. As a related aside, and further proof of the enormous breadth of the automobile system, of the four guys in my college rooming group, three of us had for most of our working life to that point been directly employed by the auto industry: I had fixed cars in California, Zack pumped gas in New Jersey, and Doug actually built cars working for GM in Michigan. The fourth, Don, painted houses in Massachusetts, but he loved to drive. He had put 13,000 miles on his VW Rabbit Diesel in a single summer driving from Boston to Anchorage and back.
What all this meant was that when I started working at Graham’s Garage in 1979, tools that previously had been essential like the lathe, band saw, welder, taps, dies, and measuring devices mostly just collected dust. On occasion, though, when for one reason or another a part couldn’t be purchased, or a design could stand some improvement, the old machines got put to work. I watched in awe as Charlie or his brother Ed stood in front of the lathe moving levers and making beautiful spirals of smoking steel spin off whatever little widget they were crafting.
Romance aside, I was enough of a pragmatist to appreciate the relevance of what I was seeing to Eli’s idea. We never charged enough to cover the true cost of making one of these special parts because people, raised like me in an interchangeable parts world, wouldn’t understand how a suspension bolt, even a hand made one, could cost 80 bucks. The bolts in the hardware store only cost a quarter. Auto repair had become part of a system of interchangeable parts and so Charlie’s custom work didn’t make economic sense anymore. In fact, to people who had no previous experience with custom work, it didn’t make any kind of sense at all. I write this without judgement implied either way; I am just describing a cultural change.
In case you have been wondering, here is where I bring this back to software. Apple recently released a 20th anniversary version of the Macintosh (remember, I’m writing in 1998). Although in the end Apple may not have as much affect on the computer industry as Henry Ford did on the auto industry, I don’t think it is a stretch to argue that the first Mac was to computers what the Model T was to automobiles. And just like Charlie Graham and the Model T, 20 years after the appearance of the first computer for the masses (any color you wanted as long as it was “platinum”) I started a company called LINC to service the needs of people who use this still relatively novel technology. At LINC we design and develop software and most of the time, just like Charlie did, we build everything by hand. We do this handwork because purchasable solutions either don’t exist or don’t meet the particular needs of our clients.
These clients love our work but are often shocked by the cost: they can get a Lexus for the same price as a pretty simple website. And quite frankly, even the best websites can’t compare to the car in terms of style, utility, performance or reliability. As a designer and a business owner this situation bothers me a great deal. LINC’s work is of the highest quality and I give my clients the best deal I can, but I am frustrated that I can’t deliver more for less. The crux of this problem is access to interchangeable parts, or in other words, the absence of a world-wide software system like the one which drives the auto industry.
It isn’t just that software development will be both cheaper and better when buying a code widget is as a simple and convenient as buying a box of screws. Not only is this obvious to anyone who gives the issue even a moment’s thought, but we are already getting close to that point. LINC’s projects often incorporate pieces of code written by someone else we download (with permission) from the Internet. The issue isn’t the parts themselves but whether they are part of an entire system that includes everything from development to distribution to repair. Ubiquitous interchangeable parts is not the end, but it is the signal that society has embraced a technology to the point that access no longer gets in the way of application.
Think about it. Other than this essay, when was the last time you heard someone talk about the miracle of auto parts stores? Probably never because for as long as you can remember they have been ubiquitous. Not only have the stores been ubiquitous, but the distribution, manufacturing and communication systems which keep them stocked with parts that actually fit into your car, and work, have been ubiquitous too. The big deal of auto parts stores is that they are absolutely no big deal.
This means that instead of worrying about the automobile, its parts and its maintenance, it becomes a tool we, as a society, count on to function smoothly. I realize that there are a lot of negatives associated with this ubiquity, but they are irrelevant to the point I am making which has to do with being able to worry about human issues rather than technological ones. The fact is that whether you approve or not, access to automobiles, and the easy living this access permits, is completely taken for granted in this society.
Software development is not like that at all. Using software to solve problems right now is mostly a painful, frustrating experience. I spend an enormous amount of time just getting simple things to work. And then I spend more time getting them to work on different systems in different places. Sometimes I move them with ftp, other times email. I write code in HTML, Java, Javascript, ActiveX, VBscript or perl depending on conditions which are too arcane to list. I buy pieces of code and spend hours decoding the documentation followed by hours working around the various bugs.
A more compelling proof of the primitive state of software development is that at LINC we design, prototype and build software. In other words, we are a consultancy that does design and production. If you think about it, you will realize that this is unusual. It is the rare architect that designs and builds houses, just as it is uncommon to find a plumber who fixes pipes and designs bathroom fixtures. The reason is that those professions have been around long enough for specialization to develop. Software hasn’t. As a result it is almost never the case that one organization specifies and designs a software application then hands it over to another organization whose sole responsibility is to create production code. It does happen, but not very often and never very smoothly.
We are in this place with software development right now because though we are beginning to see interchangeable parts we don’t have the ubiquitous software system. I have no doubt, however, that we will. I am certainly not the first to postulate this, and you may feel as though you are under a media bombardment of people telling you what the ubiquity will be like. Apparently, everything will be smart. “The computer will be the room, and the room will be the computer,” they say. There will be mobile computing, telework and hoteling. Most importantly, and most obviously, whatever it is, it will be a revolution.
I think that evolution is a more accurate description. I suspect that the experience of the future will be not all that different from what we know now. My sense is that every generation has utopian visions of the future that are reasonably correct in concept but utterly wrong in detail. Consider the car example again. Visions of the future from the early days of the car described glistening metropolises linked by shiny ribbons of smoothly flowing roadways to pristine suburbs. The reality, it turned out, is a lot more prosaic, polluted, and problematic.
This is not a criticism of technology. The point I am trying to make is that people tend to expect progress from technological change but what they actually get is just change. I am impatient for software ubiquity, whatever it looks like. This isn’t because I am interested in the technology; it’s because as a designer I live for solving people’s problems. And I know that once software development is mature, instead of wasting time messing with technology I will be able to concentrate on solving human problems. I know, though, that those human problems will no different than the ones I deal with right now.
Graham’s Garage doesn’t exist anymore. First Charlie retired, then Ed died, then Charlie died and there wasn’t anyone to run the place. Eric, Charlie’s stepson, leased it to two women who use it as an art studio. At a recent weekend of open studios I went by the place and the old engine lathe had been moved outside where it sat forlorn and rusting.
I talked to one of the new tenants and she told me how filthy the place was when she moved in. After days of cleaning and painting she had got it into decent shape. She said the floor wasn’t really suitable for an art studio (18" planks of virgin redwood that Charlie had logged and milled himself up at the Eel River) but it would have to do. The best thing was that she had finally gotten Eric to move that damn lathe out; she couldn’t understand why he had given her such a hard time about that.
I walked out without looking at her paintings, shocked, angry and sure that Charlie and Ed had done more art in that place than that rich dilettante ever would. After I calmed down I figured out a few things. First of all, that lady had no way of knowing she was talking to someone who had loved the Graham brothers. People die, things change, and as far as she was concerned she had signed a lease for an art studio, not an old garage. Second of all, I realized that I had just seen, up close and personal, the irony of technology.
Graham’s Garage got swallowed up by the inevitable forces of demographics and economics. Those forces rode quite literally on the back of the technology the garage was created to service. In 1928 the piece of property in Mill Valley that Charlie bought from his dad wasn’t worth much. After he had built the garage building it was worth a bit more, but it didn’t really pay for itself until it was the site of a thriving small business. In 1936 the Golden Gate Bridge was opened to traffic and Mill Valley was a short drive from San Francisco. This was good for business and real estate prices. Over the years the value of the property went up faster than the economic value of the business because the Graham’s never tried to grow or modernize the garage. The graphs probably crossed somewhere in the 1970’s and after that point the garage was a functional anachronism powered by the energy of two brothers and those they touched, like me. When they died there was no longer a levy holding back the economic waters and Eric could make a hell of a lot more money renting the place than he could running a garage.
The irony comes from the fact that at the start Graham’s Garage was only worth something as a place to service the cutting edge technology of the automobile. With the passage of time, the process of that same technology becoming common place created a world in which the garage was far more valuable as a place to practice painting, one of humankind’s oldest technologies. This message echoes the one I stated above, namely that technology yields change, not progress. Change is a cyclical process that has always been a part of human history. In that process today’s technology is as interchangeable with yesterday’s as the parts in Eli Whitney’s cotton gin.
