"Roman Candle" turns 20: Secrets of Elliott Smith's accidental masterpiece (slideshow)
Elliott and the friends with whom he recorded in middle school in Texas (photo courtesy of Dan Pickering)
You’d think there’d be enough laptop-toting yuppies around to fill a Starbucks in San Francisco. It’s been months since the chain equipped 50 of its stores in the area with high-speed wireless Internet access. But a tour of both Starbucks and independent coffeehouses served by the separate Surf and Sip Network uncovers a disheartening trend: Even at the spacious Brickhouse Cafi, newly renovated in the heart of Multimedia Gulch, I’m the only one logged on to a high-speed connection that costs hundreds of dollars a month to operate.
Maybe coffeehouse computing is just uncool, but a key problem is that at this point a Starbucks is one of the few places you can get wirelessly online outside of the office. Never mind that by the end of this year, more than 10 million computers will have 802.11b hardware (better known as “Wi-Fi,” for wireless fidelity) installed. The hardware may be there, but easily accessible networks connected to the Internet are not.
Opening my iBook in any of the urban parks where the locals get their afternoon sun quickly becomes an exercise in frustration. Sure, there are networks — several of them always show up on my Mac’s menu. But they’re either private nodes for nearby offices (“Rosai Group”) or inscrutable private links (“zoom0332″). Every one I try either prompts for a password I don’t have, or rejects me outright.
Techie folklore says that “war drivers” — wireless hackers who cruise the streets in cars tricked out with giant antennae and military-strength amplifiers — can surf any local Wi-Fi network at will. Maybe they can, but I can’t.
Despite the buzz over unplugged coffeehouses, free community networks and war driving, jacking in to the wireless Net is still next to impossible. Even in cities like New York, Seattle and San Francisco where public wireless projects are prevalent, working access points are rare. Technology writer Mark Durham, currently in the process of mapping all available Wi-Fi nodes in San Francisco, says you’re better off looking for a pay phone. “I’ve got about 105 listed,” he says. “But that includes Starbucks.”
If there’s one technology that doesn’t need evangelizing, it’s wireless Net access. But while there are some start-ups out there, such as EarthLink founder Sky Dayton’s Boingo, that may succeed in leading us to the promised wireless land, there are also plenty of prominent failures. Metricom’s late, lamented Ricochet network comes immediately to mind.
Wi-Fi Nation is on indefinite hold, at least until computer-carrying consumers can roam beyond the invisible tether of the base station at the office, or the AirPort in the family den. With tens of millions of customers ready to be wireless by next year, and the price of a Wi-Fi laptop dropping below $1,000, why isn’t AT&T setting up antennae for us, instead of shutting down its Digital Broadband service?
The answer is less about technology than the shifting flows of capital in the 21st century. The wireless Internet won’t be rolled out telecom-style, like DSL or cable modems. In the wake of embarrassing failures to create top-down networks, it will be built from the ground up, by a patchwork quilt of players. Imagine the gradual knitting together of cellular roaming service in the ’90s, but with 10,000 antenna owners rather than 10 giant carriers. Rather than risking billions of investors’ dollars on a ubiquitous rollout, entrepreneurs will play for smaller stakes in more proven local or niche markets: When we come, they will build it.
So how do you turn a grass-roots infrastructure into one system that customers can connect to almost anywhere? That’s where Sky Dayton comes in. The boyish founder of EarthLink, which grew from a small dial-up Internet service provider into a service second only to AOL in size, recently launched Boingo, a nationwide wireless ISP that doesn’t operate any access points. Rather, Boingo pops up its branded “Connect” button whenever any one of thousands of its individual Wi-Fi partners’ networks are in range. Customers pay a monthly fee to Boingo and the company then splits its revenue with the network owners.
Right now, Boingo has fewer than a thousand sites signed up nationwide — a piffle of the millions needed for ubiquitous coverage — but Boingo is the first high-profile consumer offering of roaming Wi-Fi service.
“Until Boingo, you really had to know what you were doing to find these networks and connect to them,” Dayton says, likening the current status quo to the early days of the dial-up Net. “We’re at that stage where you’ve got to know to set your modem to 2,400, 8 bit, and have the right serial cable.”
As Dayton learned at EarthLink, luring enough customers to build a sustainable business requires both removing those technical hurdles, and building a larger and larger “footprint” over which subscribers can connect just as if they were at home — a local call in EarthLink’s case, a one-click connection for Boingo users.
But while Dayton, a self-described “über-aggregator,” sees Wi-Fi as “the next frontier of the Internet,” Boingo still relies on someone else to invest in building the network, and that’s a hot potato in the aftermath of blowouts like Metricom and Mobilestar, which went bankrupt last year wiring Starbucks so I could have the whole place to myself.
The biggest scarecrow is Metricom, whose Richochet antennae — not Wi-Fi, but a wireless Internet service nonetheless — now sit dormant atop lampposts from San Jose, Calif., to Washington, D.C., warding off would-be imitators. One Wi-Fi entrepreneur sighs, “I think Metricom had happy customers. Just not enough of them.” A new owner hopes to restore Ricochet service late this year, but in limited high-subscriber areas.
“This is the great lesson being taught to the wireless industy,” says Bennett Kobb, author of Wireless Spectrum Finder, a reference book on U.S. bandwidth. “These vast systems like Iridium that require a global rollout before you earn one penny will fail, as opposed to these incremental systems based on local coverage.”
Setting up network infrastructures sounds like a job for the phone company, but the big wireless carriers like Sprint PCS and Cingular have a very different idea of what “wireless Internet” means than do Web-surfing Wi-Fi junkies.
“Messaging — SMS — is probably our biggest area of use,” says Steve Krom, Cingular’s vice president of business marketing and product development. “When you look at the wireless Internet specifically, what people are really looking for is the basics — news, sports, stocks and looking up very specific info related to their life.” All of which Cingular delivers, in limited fashion, via phone screens and two-way pagers that eschew normal Web access and existing e-mail accounts — you won’t be blogging from your phone anytime soon.
There are three good reasons for telecom companies to be wary of investing heavily in Wi-Fi. First, the FCC has set up 802.11b bandwidth as free, unlicensed spectrum. The telcos’ preferred strategy has long been to own the spectrum they occupy completely, rather than sharing it with every competitor who comes along. That’s what roaming agreements are for. But with Wi-Fi, jumping ship to a better deal is as easy as changing a menu setting on the new PCs and Macs. There’s no equivalent to a cellphone number you can’t take with you to discourage you from churning through providers.
Second, the telcos still have a previous legacy of spending to recoup. They’ve invested billions of dollars in third-generation wireless spectrum licenses, in order to have exclusive rights to serve high-bandwidth 3G phones — devices that fall short of what Wi-Fi offers in many ways. Verizon alone spent $8 billion on 3G spectrum rights. Writing off those costs and spending further on Wi-Fi, which doesn’t serve as a direct upgrade for their current phone subscribers, would be worse than throwing good money after bad.
The biggest roadblock, though, is the technology itself. A cellular phone antenna can reach for miles, but Wi-Fi has a range of only 100 to 300 feet at best. “It’s not carrier-ready technology,” says David Ticoll, CEO of research firm Digital4Sight in Toronto. “You’d have to put a base every couple of hunded feet, which means getting right of way [from property owners]. At any time, some number of those bases are going to be broken. It’s not an easy thing to roll out logistically on the scale that people have been talking about. And imagine if you stuck out a whole bunch of 802.11b base stations and then your customers said, ‘I really want the new 55 megabit stuff instead.’”
Wi-Fi works in the home and office because it cuts costs — if only in time wasted crawling around with cables — for the same people who install it. Get into public spaces, though, and it’s not clear who gets the benefit, beyond a few early adopters.
“The demand for wireless isn’t what you would expect,” admits Surf and Sip founder Rick Ehrlinspiel, whose six-person company manages about a hundred nodes in several cities. “That’s going to change as the price of a wireless-equipped computer drops below $1,000,” he adds — many laptops now come with built-in hardware instead of an add-on card — but Surf and Sip has had to cut staff and switch network carriers to survive until the demand grows.
With money for new hardware tight, every single access point is now open to review: Who’s paying? What will it cost? Who will get the benefit? Even the newly expanded San Francisco International Airport in the heart of Silicon Valley still doesn’t have a network. (Wi-Fi execs are wary of being quoted, but several claim a mix of internal politics and upfront payment demands crafted in the dot-com era leave aspiring providers with too little in return for the installation.)
Boingo unintentionally raised the cost-benefit issue among the free-wireless movement. Spokespeople for community wireless groups in New York, San Francisco and Seattle claim they had their members’ sites removed from Boingo’s database, since a DSL subscriber who overshares his line with the public faces a steep increase in monthly fees. “Anyone who thinks $50 to $70 a month covers 1.5 megabits per second maxed out all the time is deluding themselves,” says Mike Durkin, president of Raw Bandwidth, a DSL provider in Belmont, Calif. “Bulk bandwidth still costs us $400 to $600 per megabit per second.”
Nonetheless, the share-the-LAN crowd are doing valuable amateur market research on where the hot spots are. “My old place was near the [Highway] 101 entrance,” says science fiction author Cory Doctorow, who keeps his personal network open and lists it in an online Bay Area directory. “I got a fair number of people who needed to get their mail before getting on a plane pulling up in front of my house and hopping on my network.”
You see, despite the absence of wireless Starbucks grande latte drinkers, wireless demand does exist. Doctorow’s front yard had become a feeding trough for the very people driving Wi-Fi’s adoption: the same traveling salespeople and frequent fliers who made cellphone service a must.
“We’re at the early adopter stage,” Dayton says. “The people buying Boingo are power users and business travelers. Their alternative is a slow dial-up connection, which means being unproductive while traveling. And if someone sends them a PowerPoint attachment on a modem connection, they miss their plane.”
Glenn Fleishman, who maintains the industry-tracking 802.11b Networking News, agrees that Wi-Fi’s hot spots are along the paths beaten by frequent fliers.
“Travelers are doing personal risk management now,” he says. “They think, if I get stuck at the airport for six hours, I want to be able to get some work done. That’s worth the extra half-hour drive or cab ride to San Jose instead of San Francisco, so I can get online.” Indeed, one of the pieces of collateral that got Boingo funded was a survey that indicated 97 percent of business fliers surveyed in airports said they based their hotel and airport choices in part on Internet accessibility.
“These are the people that carry laptops when they travel — there’s 27 million of them in the U.S.,” Dayton says.
Follow those 27 million overachievers and you’ll find them connecting through Boingo’s biggest partner: Wayport, one of the few broadband success stories. The company is augmenting its previous high-speed cable jacks with Wi-Fi bases at airports including San Jose, Dallas-Fort Worth and Seattle-Tacoma. Wayport’s 4,000 or so access points are arrayed along the heavily traveled paths through airports to business hotel chains like Four Seasons, where 1.5 megabits is no longer called a luxury.
“Our partners come to us and say, ‘I’m going to lose customers over [Wi-Fi access],’” says Dan Lowden, Wayport’s vice president of marketing.
Wayport’s deal with Boingo is just a small part of its business. The company already sells its own memberships, and also partners with iPass, a provider of global employee Internet access to a couple of thousand corporations, including SAP and other giant consulting firms. For Wayport and iPass customers, Wi-Fi is one part of a larger connection toolbox, for business clients who need to get online by any means necessary, and are willing to pay for it.
Boingo is taking a different tack: It’s focusing on Wi-Fi as a separate, fledgling technology, and betting a proven name like Sky Dayton can get the modest venture capital necessary — not to build a nationwide network, but to leverage the hodgepodge of infrastructure being installed by Wayport and others. In return Boingo offers network builders a return on their capital investments, by bringing them new customers and splitting the proceeds.
If this patchwork approach sounds too flimsy to work, remember: It’s how the Internet began, and how cellphone service worked decades before Cingular arrived on the scene.
“Remember cellphones in the old days?” says Surf and Sip’s Ehrlinspiel. “You had to sign up with a different service for the next town over, because there was no roaming. And you had to carry different phones for different cities.”
Ticoll agrees the growth will be bottom-up rather than top-down. “This is going to happen as a patchwork quilt in people’s homes, offices, factories,” he says. “As it becomes economical, fun and easy, then we’ll have a mess on our hands. Classic North American wireless strategy.”
And that’s when the big phone companies will finally step in, just as they did with cellphones — making an already successful service more homogenous, seamless and mass-market friendly. Cingular’s Krom says wireless carriers may announce plans around Wi-Fi as early as this year, but don’t expect them to go rushing into the business burning big bucks.
“Most of the pioneers in this space have either gone out of business or aren’t doing very well financially,” he says. “Their business models were broken. In order to go into places like hotels, they gave way too much away from a revenue standpoint.”
Fleishman also thinks it makes sense for the telcos to wait. “Compare a few hundred Wayport locations to the number of cellular antennas in America,” he says. “It’s not even worth doing the math.”
Instead, wireless carriers are making cautious, smaller investments in safer circumstances than installing 802.11b en masse. Most recently, MobileStar’s assets have been picked up by VoiceStream, a move entrepreneurs and analysts unanimously cite as a milestone in the entry of the telcos. And part of Boingo’s $15 million first-round funding came from none other than Sprint PCS.
Dayton doesn’t doubt that if he’s successful with Boingo, the big boys will be coming around. “Right now our biggest challenge is just market inertia, and explaining the technology to people” he says, “but I’m sure in the future we’ll have plenty of competition — in fact we’ll have zillions of competitors.” And suitors, including Sprint PCS.
It might seem unfair to the scrappy crowd of Wi-Fi hackers installing their own access points around town, and to the entrepreneurs building the wireless Internet one coffeeshop — or one hotel chain — at a time. Once they reach critical mass they’ll be snapped up or steamrollered by the same telecom giants currently touting 3G phones (or 2.5G phones, or 4G phones — does it matter?). It shouldn’t be a surprise, though: That’s the way it went with cellphone start-ups in the ’80s, and with the mom-and-pop ISPs acquired by Sky Dayton in the ’90s.
Indeed, by the time I can fearlessly open my iBook anywhere in California, it’ll probably involve writing checks to SBC or Cingular. But if someone can eventually leverage the crazy quilt of Wi-Fi to let us get online from wherever we happen to be, will we really care if it’s Boingo or the Bells? Meanwhile, risk-takers from Surf and Sip to Wayport can at least take pride in the moment: While others wait around for the next big market boom, they’re unlocking smaller, safer pockets of capital to bring us the next stage of the Internet.
Elliott and the friends with whom he recorded in middle school in Texas (photo courtesy of Dan Pickering)
Heatmiser publicity shot (L-R: Tony Lash, Brandt Peterson, Neil Gust, Elliott Smith) (photo courtesy of JJ Gonson photography)
Elliott and JJ Gonson (photo courtesy of JJ Gonson photography)
"Stray" 7-inch, Cavity Search Records (photo courtesy of JJ Gonson photography)
Elliott's Hampshire College ID photo, 1987
Elliott with "Le Domino," the guitar he used on "Roman Candle" (courtesy of JJ Gonson photography)
Full "Roman Candle" record cover (courtesy of JJ Gonson photography)
Elliott goofing off in Portland (courtesy of JJ Gonson photography)
Heatmiser (L-R: Elliott Smith, Neil Gust, Tony Lash, Brandt Peterson)(courtesy of JJ Gonson photography)
The Greenhouse Sleeve -- Cassette sleeve from Murder of Crows release, 1988, with first appearance of Condor Avenue (photo courtesy of Glynnis Fawkes)