The New Building Trade

Our friend Frank Saelua is a builder. Back in the old country — in his case, Samoa — Frank taught math. I have no idea how much he knows about math, but I do know he's fully capable of building anything. Or fixing anything that's built. Put another way, Frank hacks buildings.

Frank was the foreman of the crew that built our house. This was no ordinary construction job. The architect was a 78-year old apprentice of Frank Lloyd Wright and had many of the Great Man's qualities, including the prickly perversity behind such lines as "It's the job of the architect to bankrupt the builder." While the house was a remodeling job, it was also completely original, turning a one-story ranch into a two-story modern, with cantilevered decks, whole walls of custom-made glass and almost nothing you could find in a catalog or at Home Depot.

Of course, mistakes were made, as they always are. And minds were changed. Much of the kitchen had to be redone. Pipes in the wrong places had to be moved. A bedroom wall bulged strangely and had to be flattened.

What amazed me was that Frank could look at all these problems — these walls, windows, pipes and floors — as if they were molding clay. As if nothing was a permanent structure. As if making or altering a building was merely a matter of tools and time. Need a door moved? Sure: stay out of the way and we'll do it this afternoon. Sorry about the dust.

I didn't see the similarity between hacking Linux and hacking buildings until our own chief hacker and publisher, Phil Hughes, stayed at our place over the past few days, fixing damn near everything that didn't work. This included my home-brew FM transmitter, which had baffled me for nearly a year. (And I'm not stupid — except next to guys like Phil and our 80,000 readers.) Armed with a schematic, a meter and a soldering iron, he fixed the thing in less time than it took him to tell me how much my tools sucked and what I should do to replace them. Not a whole lot different than his constant put-downs of all the editing tools that don't measure up to Vi.

Where Phil did his best work, of course, was on our Linux box, a no-name 133MHz PC clone. Using Vi and other software tools, Phil turned that thing in to a mean, clean Linux machine. When he was done, it was routing email, serving Web pages, hosting files for an office full of Macs and PCs, and serving as a desktop for reading and writing .doc, .xls and .ppt files. In other words, he made it into a real Linux system, as well as a working PC. But I submit that this is less an example of Linux doing good work than a of good worker using Linux and Unix-grade tools to do what lesser materials and tools won't allow.

Phil looked at our Macs the way Frank looked at the home tool boxes in our garage, each filled with amateur-grade implements from Sears and Orchard Supply. He made me realize that I know even less about building real computing solutions (in the literal sense of that hackneyed word) than I do about building houses. He also gave me the sense that only those of us who truly know the virtues of Vi and other Real Tools are in a position not just to solve problems but to build a better world.

For proof, look at the Internet. Much of what we know and love about the Internet — such as the way it moves mail and serves up pages — was built by guys who love to solve hard problems with good tools. These are guys who look at computing problems the way Frank Saelua looks at a bad wall.

Of course there is far more to the Internet than Sendmail and Apache. But I submit that there is something highly significant about the success of those solutions (both of which truly deserve the label) that isn't highly obvious. That's the matter of origins. There is something about where those problem-solvers came from that gave their solutions such enormous scope.

Where they came from was the Unix world.

Back in '94, when I got my first working account with an ISP, I had a hard time getting my head around all the things my stupid old Macintosh could suddenly do all at once: browsing in multiple windows; archie, gopher and telnet sessions; file transfers... even Web service, thanks to Chuck Shotten's freshly hacked WebStar. One day I was on the phone with one of the ISP geeks when he interrupted me and said, "You gotta understand: this is Unix. You can do lots of stuff at once. In fact, there's just about no limit to what you can do." At the time his business was built on cheap used Sun machines and a pile of free software. You can guess the rest of the story.

There's just about no limit to what you can do.

Combine the scope of Unix with a problem-solver's mentality and you've got a future that's equally promising and hard to see from the PC perspective. Trying to anticipate that future with PC concepts is like trying to frame a skyscraper with nothing but two-by-fours and sheetrock. And this is what Microsoft is up against right now. Whatever its ambitions, Microsoft will always come from the desktop. From the client. From one person working alone with a personal computer.

The future of computing won't be built by a company, even though we'll call it an industry. It will be built by builders and companies of builders. And both will operate on Unix-informed concepts not just of operating systems and development models, but of understanding and solving problems. And — critically, because this is a first — of doing business.

This new building trade won't be limited by one vendor's monopilistic insistence that everything be built only with their materials and tools. It won't be a made out of one vendor's pre-fab parts. Most of all, it won't be built on shaky foundations that nobody can improve because their bricks and mortar are can only be touched by their own manufacturer.

At too many companies today, even the best builders are limited by software and tools they can get only from the aisles of Microsoft Depot. This will end. Now the software business will turn into a real building trade, because the whole conversation will be liberated from hegemonistic corporate agendas by the builders themselves. Where they'll come from is the Unix world view — one where there's just about no limit to what you can do.

This new trade will be about designing, assembling, reassembling and fixing structures that are good because good professionals using good tools and materials are doing the work and learning constantly from each other about how to do it better — and doing it for the love of the work far more than the money.

Enabling this community is fundamental to everything we do at Linux Journal. Case in point: many of our writers are readers who step forward because they have useful stories to tell. Calling them freelancers doesn't cover their value. Eric Kidd puts it this way in a post at "For those of you who aren't familiar with Linux Journal, it is one of the best geek magazines currently available. In an age when Dr. Dobbs and Byte have utterly forsaken their technical roots, LJ still publishes actual source code — in some issues, well over half the articles have sidebars with program fragments. The articles are written by members of the Linux community."

So there it is. You made us what we are, and for that we owe you a hearty thanks. Now let's get back to work.

— Doc Searls

Reprinted from Linux Journal