Since it's come up in my replies, a quick rail vs. BRT thread. To start with: anyone who says rail is always better than BRT, or anyone who says BRT is always better than rail, is wrong. The best way to start is by picking the right corridor for transit, then considering mode. 
That's a question of construction cost, operating cost, level of ridership, and existing infrastructure.
Bus rapid transit usually (but not always) has lower construction costs than rail. Consider a light rail line in a city street -- if you build the same project as BRT you don't need rails, overhead wires, or electric substations.
Often, BRT also means you don't need to rebuild bridges or relocate as many utilities, because BRT is lighter.
...but: that BRT line will have less capacity. If you have a single dedicated transit lane in each direction, and traffic signal priority, you get one vehicle through per signal cycle. A BRT bus holds 120. A light rail cars carrie over 250, and it can run in 2+ car trains.
(If you have more space, to add passing lanes at stations, BRT can carry more. But that takes 6 lanes worth of space at station for lanes+platforms. And even then it's easier to stand on a train than on a bus.)
Operating costs: a bus costs less to operate for an hour than a light rail vehicle does.
...but: that bus has lower capacity. If a light rail car is full, it would take 2 buses to carry the same people. And one light rail car is cheaper to operate than 2 buses.
So the higher your ridership, the more rail makes sense. Houston's Red Line -- which runs 2 car trains every 6 minutes standing room only -- simply would not be able to provide the same speed/reliability in the same space if it were BRT.
If ridership is high enough, BRT is shortsighted and ultimately more expensive. If ridership is lower, BRT can save both operating and capital costs.
Finally, existing infrastructure.
If you're building an extension, it makes sense to keep the same mode to avoid transfers.
If you can reuse infrastructure, it can favor one mode. Converting a rail line to light rail may be less expensive than converting to busway. If you can repurpose existing traffic lanes into dedicated lanes, that will make BRT less expensive.
If the BRT lines can use existing HOV lanes, you can cover more area without increasing the cost much.
Maintenance facilities matter, too. If you already have a bus operating facility with spare capacity, but no rail line, a new rail line will require a new maintenance facility but BRT does not. That increases the cost advantage for BRT. But if you already have a rail facility...
So which makes more sense depends on what capacity you need and what infrastructure you already have.
"So wait," somebody might say, "People prefer rail." Maybe. We know that people prefer a light rail line with comfortable stations, dedicated lanes, traffic signal priority, a simple route, and great passenger information to typical local bus. There's data for that.
But I know of no data that shows that rail with all those features draws more riders than bus with all those features. What I know for sure is that putting the line in the right place makes far more difference than mode does. So we should talk more about that and less about mode.
(All of the above is about full BRT, with dedicated lanes. There are branded rapid bus routes without dedicated lanes that are called “BRT”, and which can be great projects, but I don’t consider that BRT, and it’s in no way comparable to light rail.)
(Addendum: I was being overly simplistic in terms of bus vs. rail operating costs, which depend on agency and fleet size. A light rail vehicle may cost more to operate than two buses. But the general principle holds — with enough passengers LRT gets cheaper.)
(Also: another thing that BRT can do, and rail can’t, is “Open BRT”: a BRT route with dedicated lanes and high quality stations where some or all buses continue on beyond the bus lanes as regular bus routes. That deserves its own discussion — stay tuned.)
