On August 15, 2008, among items about federal rules for special game-day service and a New York Times piece on the return of trollies to American streets, the Seattle Transit Blog made a portentous announcement to its coterie of die-hard transit buffs. “One Bus Away: the Ultimate Bus Tool,” read the headline of John Jensen’s post. Jensen went on to call the computerized real-time schedule app developed by UW grad students Brad Ferris and Kari Watkins — Watkins is now at Georgia Tech — “the new king of bus arrival prediction sites, and fills in a big gap for a strong iPhone contender. We’ve been using it for weeks, and we recommend you do the same."
But One Bus Away was not just one more piece of thumb-candy: The Puget Sound area is blessed with broad enough transit options that when one bus is belated, another offering access to your destination may be on time. As One Bus Away approaches its fifth birthday, uncounted thousands of Puget Sound transit riders have downloaded Ferris’s app to their smartphones and other portable devices.
And, as regional transit authorities gear up for their thrice-yearly service update in June, OBA is about to become the official schedule resource of systems from Pierce to Snohomish counties. Sometime this month, Sound Transit will take over management of all the data that OBA uses. Officials say they are working to make the transition seamless for users.
This is almost unalloyed good news. Ferris, snapped up out of grad school by Google, is busy helping configure his brainchild to the varied needs of cities from Chicago to Zürich, leaving the University of Washington to manage, improve and debug his software. Given its limited financial resources, the U has done well by its departed creator’s widget.
But new users will soon discover that along with the convenience of One Bus Away come some infuriating defects. OBA is a consumer interface, not a full-bore transit scheduling module. And its most infuriating defect can’t be fixed without the intervention of the outside contractor responsible for the guts of the system, INIT GmbH of Karlsruhe, Germany.
Technically, this glitch is known as “the pole problem." Riders have various printable and unprintable names for it, but usually call it “the phantom bus.” One Bus Away can only offer correct information if it’s getting accurate information from thousands of buses wandering the tri-county area. The most important bit of information is the GPS signal indicating a bus’s current location. OBA has no trouble as long as the bus is moving along its route.
But when a bus reaches the end of its line, INIT’s software gets confused. Is it still outbound? Or is it now inbound? There’s no such thing in the system as “neither,” so OBA finds itself having to guess.
The result is eerie: Your device will show that your bus is due at your stop in five minutes. Five minutes pass and no bus, so you check again. Now the bus is shown as being delayed by five minutes, then six, then eight, then 10. . . .
If you give up and start to walk to a different stop for a different bus, you will very likely hear a familiar roar as the bus you were waiting for zips away without you. A glance at your phone now tells you that the bus is on time and all’s right with the world.
In one form or another, “the pole problem” is the biggest obstacle to users as the system expands. It’s significant enough that King County Metro, the largest supplier of data for King County bus routes, is withholding payments totaling in the millions until INIT corrects it and other bugs in the software. Jeff Switzer of King County Metro says the bugs are part of the contract between Metro and INIT and that the contract funding will be supplied once all terms have been met.
Veteran users of OBA have learned to recognize the symptoms when an outbreak of phantom bus is in the offing and new users will learn in their turn. The program is just too useful to do without. And now that the resources of a major public agency are behind it, there are high hopes that today’s OBA can become the de-facto standard for further development of transit-information services everywhere.
The coordination and maintenance of OBA data will be funded in the long-term out of Sound Transit's operating budget locally. On a national level, a group of researchers from the UW, Georgia Tech and other organizations has already staffed a pilot project awaiting funding from the National Science Foundation. They plan to expand research around the project, in hopes that continued crowd-sourcing will yield the same kind of innovation that non-profit open-source software has experienced online and in smart-phone development.
