Unread alerts– why not to depend on Twitter for breaking news

In Glenview, Illinois Saturday night, a teenager pulled a disoriented women from a car that was stuck on the railroad tracks moments before an Amtrak train barrelled into it. The only reason I heard about this was because an eyewitness, David Armano, had reported via the Twitter service, which meant that he sent via his mobile phone to his network of friends; one friend forwarded it to Doc Searls, who declared on his blog Sunday morning:  "Citizen journal breaks a heroic news story."

Armano later posted to Doc's blog that he didn't consider what he did to be "real journalism." But Doc had tapped into the familiar citizen media vein. For years the deskbound netizens have been limited to only being able to happening news that they didn't need to leave their keyboards for– such as earthquakes. The first earthquake to be reported on blogs was likely in 2001; Rob Walker of Slate saw that the bloggers on the MetaFilter community had reported one in Seattle amongst themselves (he also observed that "it took just six minutes from the event for MetaFilter contributors to start dissing the press and less than 20 minutes for the earthquake to inspire a disaster prophecy for the media in general"). But that was only the blog era. Twelve years earlier, the WELL community members in the San Francisco Bay Area had reported amongst themselves the major Loma Prieta earthquake (granted, the earthquake was also beamed to the country on live television, possibly a first, as it happened just before the A's-Giants World Series game).

The key words here are "amongst themselves." In theory, yes, someone at the scene of breaking news could "break" a news story, but that doesn't guarantee it will spread very far. Armano has 424 Twitter followers, and a few of them probably got live updates to their phone or were checking IM updates that Saturday night (Armano's Twitter logs show that some of his messages were direct replies to folks). Still, it takes a bit of endurance to subscribe to the "mundate updates" of multiple friends; Armano, for example, had sent around 120 Twitter messages the previous week, most of them hyperlinks he'd found interesting.

Text messages such as the ones Twitter sends are limited to 140 characters. The narrow bandwidth may encourage on-the-scene to skimp on details that could be more naturally conveyed over the phone. In his Twitter posts, Armano only conveyed the what, but missed the details on  whothe where, pinpointing it only to Glenview) and the how. The police sent out a press release, which the Tribune stuck to Saturday night, but WBBM NewsRadio 780 dutifully got on the scene and reported the intersection (Chestnut and Lehigh) determined that it was one teenager who had performed the rescue, and they learned his name (Tommy Foust) and some insight into his conditioning (he's a part-time lifeguard). The Tribune followed-up on Monday, as did the Sun-Times and local television.

Certainly it's possible that a developing story can be made dramatic by the telegraph-like dribblings of text messages. In a recent Poynter Online article, by Mallary Jean Tenore described how the Orlando Sentinel used Twitter to pass along updates before a Shuttle launch (to tens of subscribers) For the most part, she explained, newspapers are using it as another distribution mechanism. The Oregonian of Portland has a feed with 100 followers; the New York Times has a feed with 611 followers. (Times readers have been able to sign up for email via the News Tracker service for years).

An obvious use for the the mobile phone has a place for flagging, and then learning about, public safety emergencies, Twitter isn't necessarily the magic solution. The service was down on Thursday, as Armano noted. There's no protocol within Twitter for even indicating urgency.

Here is a brief contrast of urgencies:  I was in New York last weekend with my family to see a Yankees game. My Mom would find regular updates from CNN on her Blackberry between the innings. "Guess who's resigning his Senate seat tomorrow?" (reminiscent of the update Justice Potter Stewart infamously received from the Mets-Reds playoffs of 1973: "Kranepool flies to right. Agnew resigns.") Like many interruptibles coming in over the mobile phone alerts, Senator Craig's  state of affairs didn't have much bearing on our current situation enjoying a ballgame.  Contrast that with an incident on the way to the game: while driving down the Saw Mill River Parkway towards the Bronx, we passed a car in the Northbound side that had just caught fire and effectively stopped both lanes of traffic, and could well have slowwed the Southbound traffic after we passed it. ("car on fire" is more urgent than "pants on fire.") You need a system where a driver can report an accident, it can go to emergency service, traffic bureaus, and possibly also direct to drivers. This is where a  Common Alerting Protocol comes in.

I sent a note to Art Botterell, the author of that specification, and manager of Contra Costa County's Comunity Warning System. I had first learned of his work in the wake of the Virginia Tech shootings for my article for PBS MediaShift about the FCC's efforts with the to develop a nationwide alerting standard. The FCC was directed by Congress to form the Commercial Mobile Service Alert Advisory Committee (CMSAAC), but there's been very little news on it. Botterell, who serves on the committee, told me that they are will be issuing their final report on October 12th; it may take a year or two to deploy a standard nationwide.

Broadcasting the warnings are relatively easy; getting people to report the initial report to the right place is a bit trickier. There are upwards of 200 million cellular phone userts in the United States. As for the number of Twitter users, if we take the 1990 Census Data that 1 out of 40 Americans has the surname Smith, Johnson, or Williams, and there are 3,300 users who come on searches for those last names on Twitter, and we further figure that 1 out of 6 people with those surnames live outside the U.S., we can estimate 100,000 Americans on Twitter. When I pull 40 Johnsons up, I see that only a quarter of them have more than 3 followers. Thus, the odds that somebody with a cell phone stumbling upon something newsworthy will have enough Twitter friends to be around, is something like 1 in 8,000 [These are very rough calculations; I invite any mathemetician, or even Twitter, to supply better numbers].

It's probably a safer bet for the average person to carry along the phone number or email address to send tips into a local news team. From what I've found of my local Boston news organizations, few do a good job advertising their contact points.