The question is not whether a street-level rail system works for Barcelona; it's whether such a system will work for Honolulu. On that score, the arguments for an at-grade train come up less than convincing.

Members of the Hawai'i chapter of the American Institute of Architects want the city to replace its proposed all-elevated rail system for one that runs at least partially at-grade in the urban Honolulu corridor. They argue that such a system is less expensive, easily expandable and more convenient for riders than an elevated one, assertions based on a single privately commissioned study. They also say that the changes could be made in about six months. Transit officials say no way — we'd start from square one and risk losing our place in line for precious federal funds.

Yet Gov. Linda Lingle, impressed by their ideas, sponsored an AIA forum to promote those ideas on Monday.

But neither Lingle nor the AIA have adequately countered the city's serious and compelling objections to an at-grade system. Here's why:

• Less expensive and easily expandable? That's debatable. An oft-mentioned figure is $70 million per mile for Phoenix's at-grade system, as compared to Honolulu's estimated $270 million for the elevated system. But given Honolulu's limited space and generally higher construction costs, it's unlikely the gap would be nearly that wide. Additionally, the elevated-rail estimate includes a 30 to 35 percent contingency rolled into the total cost, so the per-mile costs are likely to be lower. The city's experts also point out that to install at-grade train tracks on Honolulu's busy thoroughfares would require digging trenches up to 14 feet wide and 2 to 4 feet deep for the entire length of the rail corridor. Underground utility lines and iwi would need to be carefully handled and relocated. It's reasonable to conclude that this will drive up costs beyond $70 million per mile and take far longer than the AIA predicts.

An elevated system using 8-foot-wide pillars set 120 feet apart has a smaller footprint and more flexible installation, avoiding many of these problems.

• More convenient for riders? It's hard to see why. A train traveling down a busy street like Kapi'olani Boulevard would have plenty of company: Cars funneled into fewer lanes, conflicts with cross-street traffic and pedestrians, traffic lights, vehicles coming out of driveways and parking lots along the route. Running a train under these conditions would mean slower service, smaller cars carrying fewer passengers and a timetable at the mercy of street traffic. It's also a hazard: The Phoenix system logged 52 crashes last year, nearly half of them downtown.

• Only six months to convert to at-grade? Sounds like wishful thinking. The city argues, credibly, that just about every major feature of the system would change — ridership projections, environmental impacts, technology, routes, cost estimates — and thus require another multi-year trip through the federal review process. That could kill a rail system for O'ahu outright. Not even Lingle or the AIA members say they want that.

There's no question that an elevated rail, with a separated guideway, will do what a rail transit system is supposed to do: move large numbers of people efficiently and reliably, without exacerbating the very problem it's supposed to mitigate — traffic congestion.

If we're going to spend billions of dollars on a rail system to handle O'ahu's long-term transit needs, we should at least buy one that's going to work.