grids and the short diagonal (comment of the week)

Eric identifies an important issue for high-frequency grids, like those of Vancouver, Los Angeles, Chicago, Portland's eastside, etc … The short diagonal trip may be not much faster than walking.  Here's how he describes it, complete with clever 1980s-style computer graphics:

When discussing grids, it is important to think about trips like the following:

| (1/4 mile)
|********(1 mile)************
*****************************| (1 mile)
*****************************(1/4 mile)***

(please ignore the '*' – they exist only to make the vertical lines go in the proper place).

The grid's approach to this trip would be to walk to A, then take a bus to B, then take another bus to C, then walk to the end. However, since each bus segment is so short, even with frequent service, the waiting time still becomes a huge deal.

For example, if we assume that the buses each run every 15 minutes the expected travel time might look something like this:

Time =
5 minutes (walk to A)
+ (0-15) minutes (wait)
+ (5-10) minutes (ride bus to B)
+ (0-15) minutes (wait)
+ (5-10) minutes (ride bus C)
+ 5 minutes (walk to destination)
= (20-60) minutes

The average 40 minute travel time is just 3.75 miles per hour, equivalent to a brisk walk, while the worst-case travel time is a mere 2.5 miles per hour, equivalent to a slow walk.

With the slow speeds and huge travel-time uncertainty in the above calculations, before you even consider the possibility of bunching leading to 20-30 minute waits, if the goal is simply to get to the destination quickly and reliably, transit can't even compete with walking, let along with driving.

This relegates the use of transit for these trips to people can't walk or bike and also can't afford to drive or spend $10 on a taxi ride.

Trips like these are not edge cases. I make trips like this quite frequently. Usually, I end up either biking or jogging the entire way or walking half way and taking a one-seat ride for the other half.

My personal opinion is not that the poor handling of such trips is a failure of transit, but rather that there are certain types of trips that transit is optimized for and short L-shaped trips isn't one of them. Short L-shaped trips are simply better accomplished by some other means, such as walking, jogging, skateboarding, bicycling, or even riding a taxi, while longer trips, especially trips in a straight line, allow transit to work more efficiently.

If anybody else has opinions on the matter, I look forward to hearing them!

Eric's point connects to a bunch of intersting issues:

  • What other solution is there?   Look at the overall mobility outcome from straight, fast, frequent lines in a grid pattern, and ask:  OK, yes, this is not so convenient for the short diagonal, but what exactly can or should we do about that?  In some cities, notably Los Angeles, you'll often find little circulators that serve some of these diagonals where there is a specific market for them, such as a link between two key local activity centers.  But these are always going to be specialized because they are so much less efficient than the main grid lines.
  • Note how much this outcome depends on the overall quality of the straight grid lines.  Eric assumes they're pretty poor.  In fact, the diagonal grid trip usually has a choice of two L-shaped paths ("over and down", or "down and over") so there's an opportunity to choose the better of these two, which will the the one that uses more frequent or faster services.
  • Eric's assumptions are for a standard local-stop grid.  Frequencies are assumed to be never better than 15 minutes, and travel speed, for example, is 5-10 min to go a mile, an average speed of only 6-12 mi/hr.  Some urban lines are down in this range, but such performance should be considered a problem in urgent need of attention.  Stop spacing and a range of minor infrastructure can have large impacts, and will yield benefits that are much greater than you'll get by dissipating your service over countless little diagonal shuttles.  So there's much that can be done to improve the short-diagonal problem simply by focusing improvements on the grid lines. 

In short, I agree with Eric's conclusion. Because I tend to live in urban places where most of my trips are short, I encounter the short diagonal problem all the time.  It's a drag, but I deal with it because I'm pretty sure that it's geometrically impossible to "solve," except by undermining far larger benefits of a network that serves the whole city, and that moves fast enough to compete with cars, not with walking. 

    31 Responses to grids and the short diagonal (comment of the week)

    1. Dave July 7, 2011 at 4:43 pm #

      The other option for trips like this is to have a BIXI type system, like in Montréal. It isn’t part of the public transit (though I think it should be), but it’s precisely this type of problem that it solves for getting around (providing there’s a sufficient density of stands) and why it took off.

    2. Zoltán July 7, 2011 at 5:13 pm #

      I encounter that problem a lot too, and normally I take the longer of the one seat rides and walk the rest unless I see the other service coming.
      It seems that downtown areas are most likely by far to stimulate short trips, which is fortunate, as in any city big enough for intra-downtown travel to be an issue, bus service ought to be considerably more frequent than every 15 minutes.
      That’s unless you’re Baltimore, where the mass of buses downtown are used to provide service on about every street and on a whole lot of different routes. Hence you have a bus closer by that has a slight possibility of going closer to your destination than a straight line route would, but:
      – The poor frequency probably means you wouldn’t risk waiting for it.
      – The routes are so complex that you can’t easily remember what all of the dozens of bus routes downtown do.
      The lesson to be drawn here is that sticking to the principles of a grid and maximising the frequencies within is pretty much always the most helpful thing to do.

    3. JJJJ July 7, 2011 at 5:53 pm #

      Grids are terrible for bus transit for this very reason. Even worse is that it is unlikely that every major street has a bus line on it, meaning you may even need 2 transfers (3 buses).
      Transferring not only uses up valuable time, but can be a very uncomfortable experience. If its 34f, raining, and your transfer point is a sign (no shelter or bench) then transferring can be a terrible experience. Many might prefer a 60 minute one seat ride over a 10 (bus1)-15 (wait)-15 (bus2) trip.

    4. Jarrett at July 7, 2011 at 5:56 pm #

      Sorry, JJJJ, these hassles aren't enough to establish that "grids are terrible."  In any city with many destinations scattered everywhere, you'd need to invent an equally efficient geometric pattern that serves everyone's travel patterns, not just yours!

    5. In Brisbane July 7, 2011 at 7:56 pm #

      I have a solution!
      Problem solved!
      There seems to be this odd idea that car is illegitimate mode of transport and Public Transport should be useful for every kind of trip. It’s just not going to happen folks!

    6. Alon Levy July 7, 2011 at 8:49 pm #

      In Brisbane: your idea is why such weird routes as the 305 in LA exist. Such services are sometimes tailored to low-income constituencies that can’t afford a car. (Actually, one could argue that nearly all bus transit in the US is like that, but that’s a separate issue.)

    7. Eric July 7, 2011 at 9:46 pm #

      The problem with “just drive a car” is that cars are expensive pieces of equipment and if you can avoid the need to own one, all the money that would have been spent on car payments, maintenance, and insurance can instead be spent on more productive activities. Furthermore, if you get into town by bus, your car isn’t even available for trips within town because it’s in the wrong place. Thus, “just drive a car” means anyone making a long-distance commute into town now has to drive the entire way simply so that their car is available for short shuttle trips within the city. Again, driving a round trip of 40 miles just to have a car available for a 2-3 mile shuttle trip within town is extremely inefficient.
      I definitely agree that short, diagonal trips can often be more efficiently accomplished using some form of personal transportation, as opposed to mass transit. The trick from a planners perspective is to find a more economical way of doing this than to simply make an assumption that everybody has a car and can just drive it.
      In coming up with a solution, a good place to start is the observation that availability and directness have a much larger effect on travel time than top speed. For example, traveling at 30 mph saves a mere 2 minutes over traveling at 15 mph for a 1 mile trip. This makes all forms of human-powered transport – walking, bicycling, roller-skating, etc. cheap and effective solutions, especially bicycling, which is almost as fast as a car (maybe even faster, if parking is tight), but for a fraction of the price.
      From a planner’s perspective, encouraging human-powered transport begins by designing the streets with sidewalks, bike lanes, etc. to make human-powered transport safe and pleasant. However, bike sharing programs to make faster human-powered transport easily and spontaneously available are also important.
      For elderly or disabled people, or for groups traveling together, perhaps a bike sharing program could even be extended to include electric carts as well. Given that a Zipcar costs under $10 per hour and a golf cart should cost substantially less than a car to buy and maintain, plus take up less space to park, a 1-way rental fee of around $5 could provide service equivalent to taxi, but for a cheaper price.
      Even for people in wheelchairs, fast and efficient personal transportation can be possible with advances in technology. Already, lots of people have motorized wheelchairs. Less common is wheelchairs with powerful enough motors to go faster than a walking pace. In principle, though, there’s no reason why a motorized wheelchair can’t go 10-15 mph down bike lanes for travel around town, then slow to the 2-3 mph range for maneuvering inside buildings. While such a wheelchair might be expensive, it can’t possibly cost anywhere near what a car would cost because it wouldn’t need anywhere near the power or the bulk of a car. The fact that nobody does this seems to me to be merely an attitude of “we have cars; why bother with anything else?”.
      With a little bit of thinking outside the box, there are lots of ways to extend the car-free range of personal transportation beyond the traditional 1/4 mile walk radius. This allows transit to focus on the longer trips that transit does best, rather than getting bogged down in expensive shuttles to a handful of preferred destinations. The result is a greater relevance of transit in general and a system that is useful for a broader base of the population.

    8. Dexter Wong July 7, 2011 at 11:56 pm #

      My view on the L-shaped trip is influenced by where I live, the tropics, and if it;s not too hot I might walk the distance. But if the sun is bating down and humidity is high I would wait for a bus to take me to my destination so I won’t get overheated.

    9. alfred July 8, 2011 at 12:03 am #

      from my experience in the Bay Area…
      In SF, these short diagonal trips work well since there is a very high frequency (5 minute headways) of service.
      In the East Bay, where headways are 15-20 minutes, before I started biking I simply didn’t make those trips and tended to only go to stores that were on the same bus line that went by where I live.

    10. hallam.jon July 8, 2011 at 4:47 am #

      I actually like that walking is succesful for this trip vs. public transport. We can imagine cases – the ‘strip city’, commonly found along beaches, for example – where public transport has a bigger advantage over walking because only one line is required. However, this relative advantage comes at the cost of total mobility. You will always be able to access more destinations in the same amount of travel time with a grid, even though using a grid means you have a walking advantages to the neighbouring blocks.
      ‘In Brisbane’, using a car seems like a good solution, but car infrastructure (particularly parking) dramatically reduces density, so you pay a cost in terms of number of destinations reachable in a particular time. The car’s speed advantage makes up for this somewhat, but the more people who have one the more that speed advantages is eroded (either by delays, or by the density reduction caused by building sufficient infrastructure). Of course, if the infrastructure is already present for cars, you pay a high personal cost by choosing not to use a car, as you cannot alone create the density you’d need for it to be a good option.
      It’s also possible to imagine the case where the ‘grid’ looks like a set of stairs: One block north, one block west, one block north, and so on. In that case it’s the two-blocks north trip that requires the transfer between lines. So, essentially, we have problems with off-grid trips, whatever they may be.

    11. Tom West July 8, 2011 at 6:25 am #

      I have exactly this issue every day… I can either go east and south on the bus to get to the east end of my railway station, or walk a tiny bit east and then south to get to the west of the station. The time is about the same – but if I’m running late, I can walk faster and make the train, whereas I would miss the bus.
      There are a lot of good reasons why people don’t want to walk, even if the area is a walker’s paradise. People will be put off walking by the weather, the need to carry things with them, and the fact walking requires more energy than a bus. Those are all things that cannot be changed.
      Wider point: the travel time for any mode is equal to (access time) + (distance * mode speed) + (egress time). A mode with a slow speed can compete with higher speed mode if access/egress time for the faster mode is sufficently high and the distance is sufficiently small.
      In the example here, walking has zero access/egress time, and can compete with transit’s faster mode speed because of this and the relatively smll distance. Similarly, rail can compete with air over sufficently short distances because of the lower access/egress times.

    12. Eric Doherty July 8, 2011 at 11:24 am #

      Lets start with the tried and true – improve transit frequency and reliability towards Zurich standards (4-8 min frequency with astoundingly good reliability).
      The high tech partial solution is to put in real time digital displays so you know when to walk, and when to wait (because you don’t have Zurich-like frequency and reliability yet).
      What works for short trips also works for the start and end of longer trips. Grid-like networks are the easiest way to provided everywhere to everywhere transit service.

    13. Eric O July 8, 2011 at 2:21 pm #

      Dave said it. This is the reason for bike share (or – more to the point – bike depositing). It’s a head to head between bike-share and cabs here (and, yes, bike-taxis). Those are the superior options to beat where parking discourages short auto trips. Transit should get out of this game, IMHO, in order to diversify the mode options in the center city and create more mobility/freedom (and, hey, help support your local cab drivers!).

    14. Jarrett at July 8, 2011 at 2:34 pm #

      Eric O. Trouble is, cabbies hate short trips. Not worth their time.
      If they could they's raise the flagfall charge to discourage them.

    15. Scott July 8, 2011 at 3:34 pm #

      Two comments:
      a) Another element to consider in modeling this situation is – especially here in Vancouver, and especially at rush hour – “full bus” syndrome. The few times I actually tried to take the bus for the three-kilometre diagonal trip from home to work, I found it simply could not be done because I could not get ON the first bus despite it being “frequent.” Bus after bus was long-since “full” and passing up crowds of hopeful riders. I could either walk a kilometre to the next bus, which is Very Frequent, or simply walk the diagonal and get to work just as fast.
      b) Dave proposes a BIXI system, and Eric writes: “…a 1-way rental fee of around $5 could provide service equivalent to taxi, but for a cheaper price.”
      Something along these lines seems to be available now in central Vancouver with the advent of Car2Go, our third (!) car-sharing company. You don’t have to drop the cars off any particular place as long as it is within the service boundary. So, you could get a car for a 10-15 minute one-way trip and it would run you only $3.50 to $5.25 compared with the taxi price of about $10 with tip.
      Disclaimer: I have nothing to do with Car2Go, and in fact have myself decided to stay with the former Cooperative Auto despite them changing their name to the nauseating “modo.” in an apparent attempt to dump the hippie/granola/coop demographic and bring in lotsa lotsa hipsters. That aside, their insurance situation is far better – with a much lower deductible and the ability to totally waive it completely if you use a gold card. Further, if you don’t use the system, it only costs $1 a year to stay on the books.
      That works for me because I now use the bus only a few times a month, and car sharing every few years, preferring to deal with in-town trips – especially short diagonals – by either walking or bicycling, and longer ones by paying a few bucks to just rent a car for an entire day and not have to worry about being nickled and dimed.

    16. Daniel Howard July 8, 2011 at 4:17 pm #

      You can wrap text diagrams like the PRE tag and they line up better:

      | LIKE THIS |
      pre-formatted text!! <---+
    17. JJJJ July 8, 2011 at 4:24 pm #

      Jarrett, look at a city like Boston, the anti-grid. Point-to-point travel is extremely easy on the bus because of the way the roads are designed. You have a few major streets radiating every which way, and all the points of interest are located on such major streets.
      The problem with a grid is that there are too many major streets, so there are no clusters. 1st avenue is just as relevant as 2nd or 3rd or 4th or 5th (…) in locating your business or attraction.
      In a city like Boston, you’re limited to the few major streets that actually go places (Mass Ave, Comm Ave, beacon, Boylston, Tremont). You dont want to locate your business on a street that dead ends in 2 blocks or curves back on itself to form a circle.
      Grids = sprawl.

    18. Jarrett at July 8, 2011 at 4:42 pm #

      JJJJ.  No, disorganised grids = sprawl.  I'm not talking so much about downtown grids here as about citywide grids.  One strength of LA is that certain grid corridors are stronger than others, so they form the supergrid where subways are being built and rapid buses are currently being offered.  Corridors like Wilshire are long linear assemblies of density, so they work well with a grid pattern that moves people north-south to a point on Wilshire so that these markets contribute to the demand for bus lanes and ultimately a subway there.
      Within downtowns, 20c location practices may have made all streets seem equally important, but best practice is moving away from that toward encouraging parallel streets to specialise around different purposes — transit with strong priority on this street, cars throughput on the next street, pedestrian dominated retail street on the next.

    19. JJJJ July 8, 2011 at 9:08 pm #

      But Jarret, Los Angeles, outside of downtown, is the definition of sprawl. I disagree that some corridors are stronger than others. Olympic, Pico, Venice, Wilshire, Beverly, Santa Monica….all are superstreets luring businesses to them. I mean, whats the difference? None of them is really “superior” to the others. And that means businesses wont cluster, because if rent gets to expensive on Pico, then just move 1/2 a mile north to Olympic and you get just as much exposure. So instead of density, you get….well, you get what LA is. A nunch of 1-4 story buildings and surface lots and buses that simply cant take you from a to b without a transfer.
      That works downtown where you want a dense box, but when the city is as spread out as LA…thats not good.
      And while the best practice may be to have a car street run parallel to a pedestrian street, that boat has sailed and sunk. It’s too late, all the streets look exactly the same, and transit is diluted being forced to serve every corridor equally, and passengers are forced to wait in 95 degree heat in a bench with no shelter (does ANY stop in LA have a shelter…?)
      Contrast that with west hollywood, where geography has made it so a grid is impossible. No one will ever run a bus on those tiny meandering hill streets, and thats ok, because theyre low density….thansk to the tiny meandering hill streets! Density, and transit, can focus on the few “main” streets like Sunset.

    20. Jarrett at July 8, 2011 at 10:27 pm #

      JJJJ.  Vancouver and San Francisco are largely grid networks.  Is the City of San Francisco or the City of Vancouver sprawling?  No, they're mature cities of continuous moderate to high density, which means people are going all directions not just to downtown.  Los Angeles can grow into that if it chooses.

    21. JJJJ July 9, 2011 at 1:08 am #

      I cant speak for Vancouver, because while I have been there, it was a very long time ago.
      As for SF, it has geographic constraints (which I guess Vancouver does too). If the bay were made of land, I’d bet the grid would have sprawled all the way to Oakland, and we’d be looking at one giant suburban mass. Of course if you go south, where land is not constrained, you do get sprawling suburbs (but no longer san francisco).
      That being said, SF has the same transit problems because of the grid system that LA has.
      I hope this link works right:,-122.480564&spn=0.091341,0.181789&sll=37.754294,-122.491207&sspn=0.091341,0.181789&geocode=FZurPwId0CCz-A%3BFYp-QAIdFryy-A&mra=ltm&dirflg=r&ttype=dep&date=07%2F10%2F11&time=1:00pm&noexp=0&noal=0&sort=def&z=13&lci=transit_comp&start=0
      (I picked two random points of interest, hospital and university, at 1pm tomorrow)
      5 miles takes 56 minutes. Thats unacceptable. Even though one of the buses actually does meander around to get more directly between points. And that means an uncomfortable transfer. If SF hadnt routed a meandering bus, it would have been two transfers, and at least another 15 minutes.
      Now what happens if I pick two random points of interest in the part of the city without a good grid?
      City college and recreation center, 5 miles apart.,-122.419624&spn=0.045693,0.090895&sll=37.723139,-122.402802&sspn=0.04569,0.090895&geocode=Fca2PwIdBqy0-A%3BFTKqPwIdmoez-A&mra=ltm&dirflg=r&ttype=dep&date=07%2F10%2F11&time=1:00pm&noexp=0&noal=0&sort=def&z=14&lci=transit_comp&start=0
      Slightly less time (47 minutes) but zero transfers.
      Again, I picked this completely out of the blue (while I visit Sf often, I am not familiar with either region I picked to illustrate).
      So it looks like a confusing network of streets = faster transit times and less transfers because there are less “main” roads.

    22. JJJJ July 9, 2011 at 1:12 am #

      Oh and one other thing, grids arent good for headways.
      Say your grid has main streets 1/2 a mile apart.
      And say your city runs a bus on every single street, n/s and e/w to provide optimal coverage.
      Problem is, they will NEVER route two lines on the same street, because theyre trying to hit every avenue (which on a grid is assigned equal importance). So you may end up where every street has a single line running every 30 minutes. Or every 15 minutes. Whatever it is.
      Do you go North, then west? Or west, then north? Who knows which will be faster, it’s pure luck.
      Now say you dont have a grid system. You have a squiggly mess with choke points and curves and such. Well, multiple bus lines will be forced to run on the same streets at points because theres less main streets. So you have overlapping lines, and with it, more frequency on the overlapping bits. So now you take whichever bus comes first, even if both options mean transferring.

    23. Wad July 9, 2011 at 1:36 am #

      @JJJJ, L.A. has a couple thousand or so stops with bus shelters.
      Old style:
      New style:

    24. Wad July 9, 2011 at 1:59 am #

      @JJJJ, you also get streets with reasonably good ridership. As we know, Wilshire Boulevard has the ridership sufficient to warrant a subway.
      Most of the other streets have ridership patterns to warrant high-frequency service. In L.A., if a bus line has 10 minute service, it has the ridership to support the frequency.
      Each of those east-west streets has five-digit boarding figures. The Metro/Big Blue Bus borders of Westwood Boulevard and Crackton are for reasons of turf.
      Wilshire is the only street that has linear commercial development for most of the street. It’s mid-rise and high-rise, with some gaps between Wilton and Highland. The high-rises bend away from Wilshire southwest of Beverly Hills to Century City, and there’s another cluster of high-rises from Westwood to Embassy Row near Bundy. Interestingly, it’s in Santa Monica where Wilshire is a suburban strip-mall corridor.
      Another busy corridor is Santa Monica Boulevard. The buses run from Downtown L.A. to Silver Lake along Sunset Boulevard. This has become a famous hipster hood, and the street is mostly storefront retail. This stretch of Sunset is unusual, because Echo Park and Silver Lake are very hilly and almost all transit activity occurs along this part of Sunset, resulting in unusually high ridership.
      Off Sunset, Santa Monica serves dense East Hollywood, Los Angeles City College, and then an industrial stretch of Hollywood between Wilton and La Brea. West Hollywood begins at La Brea, and this is also a Main Street, transit-friendly and retail-rich commercial strip. West Hollywood also has a very large senior citizen population, and they keep ridership high.
      Onto Beverly Hills. Santa Monica bends northeast-southwest through here. The Beverly Hills has mid-rise and high-rise offices toward Wilshire, and a linear park away from Wilshire.
      The buses stay on Santa Monica through Century City, where the street is a freeway. Century City is on the east, while a country club with a long fence is on the west.
      Santa Monica then becomes a retail corridor, with a cluster of mid-rise offices at Sepulveda, and stays a retail corridor through Santa Monica city limits.
      The relationship between land use and ridership is quite complex, and it doesn’t factor in the transfer effects Jarrett stresses. Many of the 200-series lines (north-south, no downtown service) also have 5-digit ridership, and supply riders to these lines.

    25. BBnet3000 July 9, 2011 at 12:18 pm #

      The guy who said “just drive” is actually right in the sense that transit will rarely serve all routes to the same standard, and for some routes it may make sense to drive. This is why I am a huge fan of services like Zipcar.
      Id also like to point out that for many distances in a dense enough city, “just ride your bike” should be substituted in.

    26. Eric July 9, 2011 at 9:03 pm #

      I’m a fan of Zipcar myself. However, one of its problems is the constraint that you have to return the car to the same location you got it from.
      For example, a 3-mile-each-way trip might cost $20-25 round trip for a taxi (assuming you do it both ways), but with Zipcar, it can cost anywhere from $10 to $80, depending on how long you’re remaining at your destination before heading back. Another way to think of it is, if you amortize the cost of a Zipcar reservation based on the time the car is moving and parked, a typical trip might cost around $5-10 for travel, plus an additional $7-10 per hour to park. This is the type of cost structure that a car owner might face when driving into a super-dense area, like Manhatten.
      If there was some way to rent a car for a one-way trip and take a different car for a one-way trip back home, that would be quite interesting, as it would drastically reduce the cost for a lot of trips. However, I can also see that the logistical problems with such a system would be huge (it becomes much harder to guarantee that a car is available when you want one, or a parking space is available when you return it). But, it seems like they should be solvable.
      As a further cost-reducing measure, if the system is oriented mostly at short shuttle trips, perhaps golf carts can be used instead of full-sized cars. If you don’t need the full power of an internal combustion engine capable of going on the freeway for a 3 mile hop on local streets, why pay for it? Being able to cram 3 golf carts into 2 adjacent car-sized parking spaces could reduce costs as well.

    27. JJJJ July 9, 2011 at 10:30 pm #

      Wad: Thanks for those links. Ive only ever noticed the concrete benches (with ads) while in LA.
      Eric, I believe San Antonio has a car share system where you can leave your car wherever you want within the borders of the car share zone, which allows one way trips. They have a deal with the city in which these cars can be parked anywhere parking is legal (metered or free) and stay there as long as it wants to for free. So that helps solve that problem.

    28. Andre Lot July 11, 2011 at 1:19 pm #

      I think that, as pointed before, driving a car is just the best solution for such trips. From a technical viewpoint, a car is the most suitable vehicle for trips that are scattered in terms of origin/destination and that can’t be fit efficiently in a transit network.
      The car has some interesting technical features to make it competitive in such situations:
      – zero set-up time
      – theoretical infinite frequency (you just turn it on an go)
      – capacity to optimize its route with almost completely disregard of other trips (until there is congestion) via point-to-point, no-transfer journeys.
      As much as it is inefficient to plan mobility in a large city without some form of public transit, it is also inefficient trying to cater for every journey without considering car-mobility. In a balanced approach, they can coexist and serve different purposes. I’m amused, and also sometimes angry, at planning schemes that view car-mobility as a sort of ‘enemy’ from which cities should be ‘cleared of’.
      A (big) city where a car is required for everyone, for any journey will not be a viable place as much as a (big) city where the car can’t be used for any trip by no one. A balanced, mixed approach is needed.
      Aiming to have efficient transit reaching everyone for every possible journey everywhere in a city is bad policy as is to have any building within a maximum of a 5-minute drive from the next uncongested freeway ramp. Both are unrealistic aims which become expensive if one tries to carry them out.
      One could model, mathematically, the reach of transit versus the marginal cost of providing additional services. “Diagonal” trips will on erratic routes (irregular demand patterns, low density of travelers) certainly fall on the list of journeys whose cost of offer per user*mile would be woefully high.
      This goes not only on an urban/neighborhood scale, but also on a regional/metropolitan one – and even more. If you take an European metropolitan area with plenty of transit around, you could easily find a variety of journeys that are faster on car, if it involves going from two medium-density residential areas located on different sides of the metro area without easy, fast diametrical routes.
      It brings to question the difference between reach and viability of use of any infrastructure system, of any kind. That two places are connected by paved roads doesn’t imply that driving from origin to destination will be the fastest option. Likewise, that both places are served by transit doesn’t follow that using a complex set of routes is the fastest way of travel.
      The transit planners who demonize the car need to stop doing so.
      Now, we come to the interesting issue of car use. Though it is not the topic of this post, I think that better designed car-share schemes could do wonders for city transportation. The most impending restriction of most car-sharing programs is the fact they have fixed parking spots. If a city were to get a sufficient high number of users of car sharing schemes, it would be possible to conceive one-way rentals, which in turn makes viable to have a much higher use turnover of cars.

    29. Ted K. July 11, 2011 at 10:21 pm #

      @ JJJJ – Re : Your links …
      Your examples show why someone should do a cross-check with another service or someone who knows the territory.
      The first (SF State to Ft. Miley area) is an automatic 28-19th Ave. + 38-Geary. Both are frequent services while the 29-Sunset is a secondary that can be erratic. Also, the 38L is an option if you don’t mind walking a couple of blocks. The 38 is direct to the hospital in daylight hours*. [IIRC the 2-Clement, a parallel route to the 38-Geary, used to run out into the Forties but was shortened quite a while back.]
      * See Ft. Miley service section –
      The second (Rec. Ctr. to CCSF) is a hoot because the rational routing, 44-O’Sh. + 49-M-VN (transfer at Mission + Silver) isn’t even listed. Also, listing the horseshoe (K/T via downtown SF) should only happen when one requests the WORST route possible. The one-seater, 54-Felton, is a community service route that often has missed runs.
      Is SFMuni trouble free in its coverage ? FRACK NO ! Like a lot of systems it is slow when going from outlier to outlier. But when traveling from outlier to downtown it’s pretty decent due to limited stop service in certain corridors. And sometimes it pays to go around two sides of a triangle even if it takes more time on the bus but less time waiting.

    30. Eugene Wong August 31, 2011 at 9:03 pm #

      Maybe the trick is to look at it like elevators. In the olden days, they used to have elevator operators. Certainly, there must have been ways to reduce inefficiencies.
      Maybe the trick is to have a shuttle run with a simple vehicle that doesn’t really follow a schedule, but tries to meet a certain frequency. Since it’s not a bus, you wouldn’t need special training.
      For example, every 5 – 10 minutes, send a van from A to C. The van could leave as soon as 1 person gets in, just as if it were a taxi. With such a short route, they could avoid bunching up by going back to the start. The van is shorter than a bus, so it could easily turn around in a parking lot.

    31. Jarrett at August 31, 2011 at 9:38 pm #

      Welcome to the developing world, Eugene! Where labor is cheap, many
      transit services work much as you describe.