E Line F Line
Transit Signal Priority (TSP)

$5.97M (16%)

$9.93M (28%)

Passenger Facilities
(Shelters, lighting, etc.)

$5.1M (14%)

$7.64M (22%)

Real Time Information Signs

$790,000 (2%)

$800,000 (2%)

New Buses

$24.9M (68%)

$16.8M (47%)




Following last weeks announcement of FTA’s Very Small Starts funding of the E and F lines I started wondering about the cost breakdown of RapidRide. I have some answers thanks to Rochelle Ogershok at King County Metro. For each line, the total cost is roughly the same, but the breakdown differs significantly between the two. For the E Line, almost 70% of capital costs go towards new coaches, which certainly reflects the longer cycle time, and thus larger number of buses necessary to achieve 10-minute peak period headways.

Comparing the two lines, the extra cost of new buses on the E Line is offset through lower roadway/communication/TSP costs and passenger facilities. This is likely because SDOT and Shoreline have already done most of the work to prepare signals for TSP along the corridor by upgrading signal controllers and cabinets, which is necessary for various component of the RapidRide ITS system to operate. The lower relative cost could also be related to the need for fewer concrete bus stop pads along the E Line.

Despite the larger number of “stations” on the E Line, passenger facility costs are higher for the F Line. The only reason I could see explaining this is better and thus cheaper access to power along Aurora, less right of way purchase necessary to accommodate stations and fewer sidewalk improvements for the E Line compared to the more suburban F Line. The smallest part of the budgets, just 2% for both routes, is spent on real-time information signs, although things like power and communication, which are included in the other cost categories, are need for real-time information signs are to operate.

My personal take on these numbers, which I have said before, is that American BRT over-emphasizes slick-looking buses at the expense of more broadly investing in speed and reliability improvements that benefit frequent bus routes, regardless of branding. That FTA specifically funds projects which meet its “BRT” criteria, without setting caps on the the percentage of a project budget that can go to buses, exacerbates this problem, and could create an undesirable incentives for transit agencies to see FTA BRT funding as free money for bus replacement, rather than a challenge from the FTA to improve the quality of bus transit.

57 Replies to “RapidRide E and F Cost Breakdown”

  1. “…which certainly reflects the longer cycle time, and thus larger number of buses necessary to achieve 10-minute peak period headways.”

    It’s too bad the E line is going to 10-minute peak headway since the service that is there now runs about every 7 minutes.

    1. But most of the day it isn’t 7 minutes. I’ve never ridden it when it was 7, I usually just calculate 15. Perhaps a reliable 10 minutes would be great. The 358 comes by every 15 minutes, then 22 minutes then there will be three of them back to back and there won’t be another one for 30 minutes.

    1. RapidRide E’s operating cost will be more than SWIFT’s, due to shorter headways on Saturday, service on Sunday, and better-paid operators.

      Even at peak, SWIFT-to-E transferers can look forward to planning a 10-minute layover at Aurora Village, as I have heard nothing about the buses trying to time things to exchange passengers.

      But that reminds me: Has there been any effort to divert 511’s routing to meet up with SWIFT? It seems such a painful miss that there is no express route that meets up with SWIFT (except at the far northern terminus) off-peak.

      The 510 could take over picking up at Ash Way, and have the 511 make the short jaunt from Lynnwood TC to Hwy 99. Connectivity, please!

      1. Thanks for attention to the 511 and its territory, where my average workday puts me aboard pm rush hour service.

        My current destination is about a mile north of Lynnwood Transit Center on 44th Ave-though occasionally also work along SR 99 a half mile north of 196th St.

        That time of day, I wouldn’t waste express schedule between the Transit Center and SR 99. Traffic can hold a schedule five minutes or more just getting out the north end of the center.

        Recalling five years ago when present work schedule developed, what’s really needed is for first-world local transit service to be restored to Snohomish County.

        At a time when current Administration desperately needs some visible accomplishments, a cheap and easy one would be to very publicly waive whatever rules forbid SWIFT from covering for the old CT 101, which really worked.

        CT’s remaining operations are spread so thin that I can’t even use the new schedules which would work very well for me if they were runnable.

        But your point about non-communication between agencies is bold-faced underlined italicized spot on. As both a passenger and a transit driver, my blood boils every time a driver on any service tells me I have to miss a connection because, well, “they’re a different agency.”

        Maybe we the voters need to fix that. We live and work in a region. We need to ride as we live.

        Mark Dublin

      2. @Brent – I suppose this idea maybe could work, provided you started with the premise that you have to have separate buses going down I-5 for Lynnwood and Everett. But if we did this, it would be virtually impossible to combine the 510 and 511 into the 512 because it would add too much time to trips to Ash Way and Everett.

        Given the choice between a better connection between 511 and Swift vs. doubling the frequency between Lynnwood and Seattle, I would unquestionably choose the latter.

        The ideal solution to the bus-switching delays at Aurora Village would be for the E-line and Swift to someday thru-route each other and go straight down Aurora the whole way. In practice, the E is going to have to become much more reliable than the 358 in order for this approach to become practical. But maybe, someday that can happen. If we can just find a way to get the E line through downtown onto Aurora quickly and reliable, once it’s on Aurora and moving, the rest of the route should be pretty reliable.

  2. I looking forward to the next Route Performance Report, should be out now or soon, to compare RR-A after a couple of years. They doubled the frequency on the old MT174, hours are way up ??, and riders are up too ??. All this money should be paying dividends by now in better riders per (pick one), or some serious head scratching should be going on.

  3. I assumed that rapidride e would run at a higher frequency during peak periods. (See as the 358 already does so). Are we sure it will only run 10 minute headways?

    1. I think there’s two things going on here. First, even if the 358 nominally operates 7-8 minutes during the busiest part of the peak period, everybody knows that the buses are going to get bunched and the real headways the actual customer would experience is probably at least 10 minutes anyway. Hopefully, the E-line can maintain its headway more reliably, for example, by selectively applying signal priority to buses running behind schedule, but not to buses running ahead of schedule, or by having buses start skipping stops when no one is getting off and the next bus behind it becomes sufficiently close.

      Second, my past experience with RapidRide is that Metro prefers to err on the side of under-promising with respect to peak headways. I’ve ridden the B-line a few times during the afternoon peak and even though the paper headway was 10 minutes, the actual wait time I experienced was quite a bit less than this. And based on the number of buses I saw go by in the opposite direction, real headways were probably more like 7.5 minutes than 10.

      The under-promising actually makes quite a bit of sense. Under-promising, in general, tends to make happier customers – think about airlines and Amtrak pad their schedules to increase their on-time arrival rates. Reducing the promised peak headways means buying extra buses that only get used during the peak, but sit there empty all other times. Because buses sometimes have to get taken out of service for repair, it is necessary to have more buses in stalk than what you actually need for service on a given day. When buses aren’t taken out of service for repair, you can quietly put these extra buses into service during the busiest times and make riders happier. But, even when it’s not possible to do this, you are still fulfilling your promises to your riders.

      1. “by selectively applying signal priority to buses running behind schedule”

        That’s not part of A or B that I’ve heard of. TSP is just on all the time and is designed to move buses through the corridor as quickly as possible. I’ve never heard of any logic that looks at how far ahead or behind schedule we are.

        “my past experience with RapidRide is that Metro prefers to err on the side of under-promising with respect to peak headways”

        Correct… I and my follower were delivering 10 minute headways well past the advertised 9am (almost until 10am) for most of the line. I suspect they start to get more buses back out in the afternoon before 3pm as well.

    2. Metro has to adjust its calculations due to the effect on ridership of the RFA transition. Ridership within the former RFA will drop, and perhaps noticeably. That can be figured out from the studies of how making whole bus systems free caused ridership spikes, that then went away with reimplementation of fares.

      On top of that, Metro has to run fewer buses through downtown in order to keep the buses moving.

      Expect headways on various lines to increase, not just from unreliability downtown, but also from Metro right-sizing capacity on each line in response to intra-downtown peakloads getting smaller.

      1. Yeah, the unreliability of downtown is going to be one big difference between the A/B lines and the C/D/E lines. If a bus has 15 minute headways and you can only predict how long it will take to get through downtown to the nearest 15 minutes, everyone getting on the bus after downtown is going to be screwed, at least for outbound trips.

        I’m not too concerned about right-sizing capacity in response to intra-downtown peakloads getting smaller. The buses should be at their maximum load just before entering downtown and just after exiting downtown, not within downtown itself, where many people have already gotten off (for inbound trips), or many have not yet gotten on (for outbound trips).

      2. Ridership within the former RFA will drop, and perhaps noticeably.

        I am still quite skeptical of this. I think very few riders are currently making intra-downtown trips without a valid transfer – the RFA is just so small. I think Metro said they expect a couple percent drop in downtown ridership, but even that seems like much to me.

        If ridership drops significantly from the RFA elimination, it’ll be because of new reliability issues, not the fact that you now have to show a transfer to go from Westlake to the ID.

      3. I don’t think reliability will hurt intra-downtown trips. It’s called a River of Frequency for a reason.

      4. I think the Rapid Ride will have an advantage here, because the buses can load all doors for Orca users. Rapid Ride may be the only on time service out of downtown.

    3. I could be worn about the ten minute peak frequency. I looked quickly but couldnt find any information to clarify it to me.

      1. The 140 is scheduled at 40 minutes end-to-end. The 358 is scheduled just under an hour. I’m afraid you are right about the headways.

        I think Metro will figure out how to get the buses moving downtown by the time the E Line opens (fall 2013). Right now, they are just in denial that there will be a problem, but they have plenty of tools they just haven’t had the political will to use.

  4. Very good point about stress on the appearance of the buses over the far more necessary elements of lane reservation and right of way. It also doesn’t help any that in some ways the buses on the SWIFT and our own BRT lines provide worse travel quality than regular buses.

    It’s a fair argument that these buses are designed for more comfortable standing loads. But most if not all of the seats seem small and badly-positioned. And for all the glass, sight-lines from inside the bus seem to consist mostly of pillars between windows and thick black frames.

    Granted, the art of low-floor coachbuilding is still very young. Above points could be chalked up to experience aboard a bus that actually spent most of its trip-time in motion. But it’s really infuriating, especially on a freezing winter day, to watch a SWIFT bus approach your station behind traffic turning into a mile of driveways- and stop at a red light twenty feet short of your zone like any ACCESS van.

    The whole idea of “branding” in current usage begs the question of who is the cow. It contains the implied insult that transit money is better spent on PR than operations. Maybe the average person in this country really is too dumb to notice- young people have been known to kill each other over crap made in China that’s got the current mandatory label.

    Like most of what’s wrong with transit in Seattle and the rest of this region, the real problem is that lane reservation and signal priority take some political decisiveness and guts. Whose cure requires very large numbers of the general public, and a fair number of job-creating employers to get in the faces of elected politicians and insist.

    Because in this economy, no matter how impressed anybody may be with streamlining, the average person’s boss still knows enough to replace somebody fresh off a late-running branded bus with someone who drives a car, regardless of the brand on its fender.

    Mark Dublin

    1. The lack of political will for ROW comes from the state and cities (including, alas, Seattle), not the agencies.

      1. Notice I said “elected politicians”. Many of KC Metro and ST staff and I have worked together since we wore the same brown uniforms.

        Only good thing separation brought me- leaving out seventeen years’ better physical health- is that King County Council Members and Sound Transit Board Members are in my chain of command, rather than the other way around as is the case with staff.

        Fact that something is “a political decision” doesn’t mean it’s something I can’t change. It means it’s points against me, and a letdown of my friends on staff, if I don’t do my damnedest.

        Mark Dublin

    2. One of the problems Metro is facing is how to have enough width for the passive restraint slots. At least, that’s what I heard a few months ago. They don’t know how to make them fit on the new Orions (the ones with the Link-style metal seats) with the seat configuration they’ve installed.

      If there is a tradeoff between seating and slots, I’ll gladly give up a couple seats so that someone in a wheelchair can ride “independently” (if they so choose). I really do want passive restraint slots to be part of any and all fleet replacement.

      1. Right on that too, Brent- for Tunnel service even more than BRT.

        Mark Dublin

  5. Do the figures include what cities are budgeting for capital costs on these lines?

    Also, with more buses working the E Line, the operating costs of that line will be proportionally higher than that for the F Line, FWIW.

    With a little investment in 24/7 fare enforcement (shared among the alpha lines, if necessary), hopefully the pay-at-the-front-between-7-pm-and-6-am confusion can go away, so that riders will more consistently take advantage of off-board payment, and so can the willingness of some dumb passengers to assault operators or other passengers. Safety at stations/stops is something that still needs a lot more thought and work, though having fewer stops helps in that regard, assuming the least safe spots aren’t selected.

    1. I’m pretty sure the costs shown only include work being done now to get RapidRide up and running. So for example SDOT has done a lot of signal upgrade work on Aurora and that isn’t included in the above project costs.

  6. Can anyone explain why there isn’t a rapidride line planned to replace the 550?

      1. Right. For the next 11+ years, there is a huge gap in the system.
        You can go from Bellevue to Redmond via RR, and from Seattle to points north and south, but nada for options from Eastside to/from Seattle.

        The mesh isn’t connected.

      2. @PSF, I think you are overrating the travel-time difference and headway difference between the 550 and East Link.

        The 550 still is much more express, and has fewer stops, than the RapidRide lines, and has higher ORCA usage to boot. A full list of the stops is in the back of the ST route guide. It also has competitive or better headway most of the time except weekends.

        Also, a few more trips got added to the 550, starting Monday. ST is willing to add more trips to alleviate overcrowding, at least on this route.

      3. “The 550 still is much more express, and has fewer stops, than the RapidRide lines, and has higher ORCA usage to boot”

        Higher ORCA usage, yes. Off-bus payment, no. As much as I’ve griped about cash payment on RR B, a large, and growing, percentage of passengers pay off the bus. Loading at Crossroads and 140th are especially good examples of the potential. The 550, on the other hand, has several trouble spots where the bus will sit and wait as people pay their fare. Even with almost 100% ORCA usage, South Bellevue and Mercer Island P&R can take a while to load.

    1. The 550 is a Sound Transit route, not a Metro route, and has all the characteristics of RapidRide already except for the off-bus fare payment.

      RapidRide is a Metro program, not a Sound Transit one.

      1. Right – our seamless transit system should have randomly different policies and fares and services because there are two separate agencies with different viewpoints.

        That’s my attempt at sarcasm. If we really had a seamless system, we’d align policies, fares, and procedures to the greatest extent possible, and if there were a Rapid brand, we’d have it across agencies. And consistent fares and transfer policies.

        Frankly, the Rapid approach should be used across all high-frequency routes (I mean stop diets, signal priority, off-board fare payment, etc.) I don’t know that the trend toward lettered routes is hugely helpful, however, it’s a little confusing. And prohibit customers using 10-min or 15-min service from accessing the schedule is also a rider-unfriendly policy.

  7. The more I read about the RapidRide, the more disappointed I am. In general, I think it has its place. Quick boarding along with fewer stops and great frequency can make for a great system. But here are my gripes:

    There is no real schedule. They only list the departure times when it is more frequent than 15 minutes. This means you have to have to guess when it will arrive if you ride it during the day. I know the idea is that it runs frequent enough to not need a schedule, but 10 minutes (let alone 15) doesn’t cut it. Get the frequency down to four minutes and we can do away with the schedule. Strangely enough, Metro has a timetable for normal buses that travel that frequently.

    It really isn’t that rapid. It takes 35 minutes to get from Tukwila to Federal Way. I know it isn’t meant to be an express, but maybe that’s the problem. We could really use the external fair system on a lot of the express buses. The 41 is a fast bus, but it takes a huge chunk of its time at Northgate, letting people on or off.

    There may be no practical way to solve this problem, but it is very disappointing to see that the first stop on Aurora for the new E Line is 45th. If they added a stop on 39th or so, it could make it reasonable for someone to take the bus to Fremont. So basically they are throwing away a huge chunk of potential riders so that they can rapidly make their way to 85th, where it then makes stops every five blocks or so. In other words, for much of where the Rapid Ride travels, it makes a lot more sense to just drive. This is very disappointing.

    1. It’s not quite as bad as you say. First, even though Metro doesn’t provide a schedule when the bus is running every 15 minutes, OneBusAway still provides real-time arrival info (they didn’t use to when the service first started, but they do now). Furthermore, with the not-too-great reliability buses have today, even at 15 minute headways, a schedule doesn’t do that much good. For instance, if the schedule leads you to arrive at the bus stop 5 minutes early for a bus that actually arrives 5 minutes late, you experience a 10 minute wait for the bus, which is on par with what you would expect if you just show up at the bus stop without looking at the schedule. Real-time arrival info does allow you to do better than this and, yes, the RapidRide routes do support this, along with all other bus routes.

      Also, the statement that the first stop on Aurora for the new E line is 45th is not correct. There is a stop at Aurora and Galer, serving the eastern part of Queen Anne. Yes, a stop for Fremont right above the troll would be nice in an ideal world and, maybe if enough people clamor for it, it could be added in the future, but, for now, the cost of elevators to move people between the bus stop and Fremont proper would be quite expensive.

      I do agree that the biggest problem with RapidRide is that it really isn’t that rapid. The stop spacing in Shoreline is a bit closer than I would have liked, and historically, when confronted with choices between keeping the bus on a straight line and deviating to serve whatever special interest demands special service at the expense of everyone else’s time, Metro has chosen to deviate. The B-line’s deviation to 152nd Ave is absolutely rediculous. I work in a Microsoft building right at the corner of 156th Ave and 31st St. Even though the B-line has a stop right across the street, it’s actually faster to walk not one but two stops down and catch the bus at 24th St., avoiding the deviation to 152nd St. and the lights and roundabout traffic that come with it. As a general rule of thumb, any time it is faster to walk to a stop further down the line than catch the bus at the closest stop to where you started from, the bus route is suspect – buses should be running fast enough that walking down the line to catch the same bus should never make sense if the goal is to save time.

      A similar to-deviate-or-not-to-deviate debate is on with the E-line around Phinney. While I’m hoping Metro keeps the bus running straight down Aurora, I’ve resigned myself along time ago that it’s going to keep on deviating to Linden just like the 358 does because a few people complaining about walking one extra block and waiting for the light to cross Aurora trumps the time of everybody riding the bus to stops further down the line.

      1. One thing that I can’t quite understand is that the stops are about every 10 blocks in Seattle where there are no sidewalks, but in Shoreline where there is great walkability, the stops are every five blocks.

      2. I don’t think it is necessary to build an elevator to connect Fremont to the E Line, just a a bus stop. An elevator would be nice, of course, but there isn’t one on 45th either. I think you could add a stop right after the Bridge street exit (the equivalent of about 37th). There is a crosswalk across Bridge street, as well as stairs connecting both sides of Aurora to 36th. That puts you right at the troll, a short walk to the heart of Fremont and about 8 blocks less walking than 45th.

        As to the lack of a schedule, I don’t think OneBusAway is a good substitute. If I plan on taking a bus from the beginning of a buses route, I can be reasonably assured that the bus will arrive at that time. You are right, I will allow an extra five minutes (in case the bus is early). But more than anything, I know it will be that bus, and that the bus will take (roughly) the same amount of time to get to my destination. This makes planning much easier. I do this all the time. If I have to be downtown at 9:00, then I’ll look at the schedule for the 41 and work backwards. I can compare it to other buses that go downtown. I can’t really do that with RapidRide. I have to guess how long it will take and then allow a full 15 minutes because I don’t know when it is expected.

        If this really is the plan, then it is pathetic. Basically, RapidRide is saying “We aren’t really sure when the buses will arrive, but you should see one every fifteen minutes or so and we aren’t sure how long it will take to get to your stop, but it takes 35 minutes to run the entire route at 3 in the morning, so maybe you can figure it out.”

        This isn’t terrible, mind you, just worse than a regular bus. That is the really disappointing part. RapidRide could be something special and something great. There are areas of the city where it makes a lot of sense. But it needs to run way more often, way faster and they need to print out expected travel times. As it is, I think of this as just another bus, and not that great of one.

      3. Regarding the Aurora vs. Linden option around 65th – 70th Sts., a bus stop along Aurora at this area would be terrible for riders especially in the northbound direction. Relying on a signalized crosswalk and having to cross Aurora Ave does not make for a pleasant pedestrian environment. At least two and more likely three stops for the #358 along Linden/Winona are being consolidated into one stop for RR-E. Fewer stops will speed up the service relative to #358, and many more potential riders are within a 1/4 mi or 1/2 mi walkshed of the lone stop along Linden than would be the case for an equivalent stop on Aurora.
        And no, I don’t live near this stop on Linden but I do sometimes ride #358 north and don’t mind this relatively small deviation.

    2. Ending pay-as-you-leave will speed up many of the outbound routes like the 41 (at the cost of downtown speed and unreliability if Metro does nothing additional).

      But there is an additional tool that could be employed to help speed up high-volume boardings at transfer centers and stations: Loading assistants armed with ORCA readers, working the back doors.

    3. Has the routing and stops for RR E been finalized yet? There was the public meeting last year about Linden but I never heard anything afterward.

      In Shoreline, Metro put the stations where the City of Shoreline asked it to. Shoreline has promised TOD around every station. (And Lynnwood has also zoned for TOD around Swift stations.) Shoreline has BAT lanes for the route, not 100% coverage but it’s approaching it. All this is in stellar contrast to what Seattle is doing south of 145th, which is unbecoming of a city.

      I don’t think you can realistically get down from a station at 39th without elevators. I have asked Metro a couple times for an elevator station there, but I’m sure Metro just laughs at it given Metro’s budget situation. We seriously need a levy to bring all core routes up to 15-minute frequency, restore the standby service that fills in when a route gets off schedule, and to make station-specific enhancements like this 39th elevator.

  8. Does the Rapid Ride program represent an increase in Metro’s total fleet size?

    Or is there a 1-for-1 retiring of an older bus for each red coach put on the road?

    If so, then “Rapid Ride” is primarily a way for Metro to get access to federal funds for bus replacement.

    1. Metro has limited base space to park buses. They have little choice but to auction off the surplus clunkers.

    2. There will always be equipment displacement, upgrading a line to a BRT line with new equipment will usually render the old equipment surplus, unless theres a general service expansion at the same time. Replacing a line that uses 6 buses with a line that uses 12 new buses, means the 6 old buses are now excess, and can be shifted around to retire 6 of the oldest, or otherwise life expired buses (based on mechanical problems, mileage, accident, etc.) you cant keep everything.

      1. You would think that King County has land someplace where they could mothball buses that still had useful life. Because this county is going to add upwards of a million people in the next decades e.g. the lions share of the region’s forecasted growth. It would make sense to have some reserve to accomodate expansion without having to spend additional capital to expand services. I think a reserve of about 100-200 buses might do it.

      2. Metro has a history of pushing equipment beyond it’s planned service life. The Breda trolley buses being the poster child for this. With budget constraints capital equipment took another hit in order to save service hours. One thing KC Metro is not short of is buses that need to be gotten rid of. Besides, equipment does not like to be parked without use. Maybe if there was a big lot in the desert but not in the Pacific Northwet.

    3. “If so, then “Rapid Ride” is primarily a way for Metro to get access to federal funds for bus replacement”

      I wouldn’t say that at all. I’ve griped about RapidRide being oversold but there are very real improvements in speed, comfort, and (likely) ridership. There is still work to be done but I’d hardly call RR some sort of scam.

  9. Is it just me, or is everything about the RapidRide stops just terrible looking? I don’t mind the busses as much, but the new “branding” that Metro has come up with for the stops they are installing all along third just looks flimsy and cheap. What is with the little red flags sticking up in the air? And it’s hardly enclosed at all. Metro needs to take a serious look at their shelter design and overall graphic identity anyway. I know this is a sideways rant, but looking at these stops every day (as a 358 rider in anticipation of the E line) is really making me think of how much Metro could use a brand overhaul in general.

  10. E Line will cost $36.74 total? Nod bad, we just need 8 people to toss in about 5 bucks each.

  11. It does seem to me like Rapid Ride is primarily just a way to get some fancy buses on the federal dime. For the amount of money being spent, the minimal decrease in my travel time between the current 358 and Rapid Ride just isn’t worth it. I mean, why invest all this cash when the service will be pretty much the same? I hate that it takes me almost an hour to get home to Bitter Lake from the financial district–it’s about the same amount of time as when I lived in south Snohomish County. And my bus to Edmonds wasn’t filled with crack addicts.

  12. Wish ST 522 would make just ONE additional stop: near 80th and 15 Avenue NE to compensate for the terrible Metro evening and weekend service on the 72 and 73.

    1. Or even just have it stop at the stops used by the 306/309/312 at 85th & Lake City Way.

  13. I count 32 stations/stops on RR E. If I use a guess of 100 sq-ft of sheltered area at each that works out to $1,600 per sq-ft. Where’s the granite counter tops, SS appliances and media room? For the $12.7M spent on “shelter” for these two lines (where there already were bus shelters) we could afford a bed for every homeless person in Seattle, sans granite counters of course.

Comments are closed.