The problems with buses are all caused by cars. Why not build new roads solely for the use of buses? This would allow you to superimpose a subway like system on top of city's without the restrictions and cost of rail. Let specially licensed freight use it also to give business benefits.
Sorry, but that's just not true. In a relatively small city with narrow streets like Cambridge, buses have just as many problems navigating around cyclists, awkward intersections, and indeed other buses. And the problems buses cause, particularly when only sparsely occupied, have nothing to do with cars at all.
> Why not build new roads solely for the use of buses?
Where are we going to put those?
They're trying to create new roads only for the use of buses, by taking away lanes that used to be for everyone and designating them to be bus lanes instead. That is making things much worse for everyone else, because it means 50% of the road surface on major arterial routes is unused most of the time. And it isn't even helping buses that much, because they still have to get to and from those major routes on the back roads they serve, and they still have to stop at lights or give way at roundabouts just like everyone else.
(Cambridge also has far too many sets of traffic lights for its size, which introduces an entirely artificial ceiling on the capacity of its road network for reasons no-one seems to quite understand. You can usually tell when the lights fail at a major intersection, because the drivers all start being polite and taking turns to proceed, there are hardly any queues in the area, and journeys are invariably quicker and more pleasant. Note to people wanting saner traffic systems for cities in future: competitive routes and stop/go lights are an evil that will cripple your efficiency.)
> (Cambridge also has far too many sets of traffic lights for its size, which introduces an entirely artificial ceiling on the capacity of its road network for reasons no-one seems to quite understand. You can usually tell when the lights fail at a major intersection, because the drivers all start being polite and taking turns to proceed, there are hardly any queues in the area, and journeys are invariably quicker and more pleasant. Note to people wanting saner traffic systems for cities in future: competitive routes and stop/go lights are an evil that will cripple your efficiency.)
Incidentally, I did hear about a study that indicates that roundabouts are far better for safety than traffic lights. Unfortunately, I don't remember the reason they concluded for this. Suffice to say... I don't really see much reason to ever use traffic lights if they can be avoided.
Only if you're in a car. For cyclists they're far more dangerous than traffic lights (where you can do a hook turn), and also pedestrians (you can install pedestrian lights, and the cars will be going slower anyway).
It depends a lot on the design of the roundabout. Our traditional medium-sized model here in the UK, with two or three lanes coming in as one block from feeder roads, certainly isn't great for cyclists. However, we seem to be moving towards a more continental style, where for a medium-sized four-road roundabout you'd have one lane peeling off to the "first exit" and a physically separated lane onto the roundabout for those going further round.
This is generally an improvement for everyone, because you only ever have to worry about merging with traffic from one other lane at once. It's also an improvement for cyclists in particular, because you can construct those single lanes so that either they are wide enough to pass a cyclist with a good clearance or they are narrow enough that overtaking is clearly not an option. When there are several adjacent lanes, an aggressive driver will often force a cyclist into the next lane over (even if there's another vehicle there). When doing so requires driving through solid concrete lane dividers, funnily enough it doesn't happen so often...
It doesn't help when making right hand turns* or going straight. In that case, a cyclist has to cut in front of two lanes of traffic (left and straight/right), which isn't very good for either them or the cars.
You also have to consider that the cyclist needs to give way to the traffic from the right. If they've stopped, they need an extra large gap in the cross traffic, since acceleration on a bike isn't quite as good.
* - This is for areas where you drive on the left, obviously.
The risks you describe are all true, but mostly apply at traffic lights as well: a cyclist still has to navigate into the right-hand lane at a junction if turning right from a two-lane (or more) entry road.
In practice, modern design practice for roundabouts ends any separate cycle lanes well before entrances to the roundabout, thus allowing everyone to merge on approach in plenty of time. Also, the roundabout itself should be designed to limit both the number of potential points of conflict and the speed of traffic crossing those points if shared use is expected, for example by the use of solid islands and dedicated lanes that must be chosen on approach and then followed throughout in a predetermined spiral path around the roundabout. That means cycling on such roads is usually both more efficint and safer than cycling through signal-controlled crossroads of a similar scale.
For larger roundabouts, cycling around with several lanes of traffic is rarely advisable anyway and alternative provisions will need to be made. Although dedicated cycle routes like underpasses come with their own safety concerns, particularly at night and not necessarily anything to do with traffic, most roundabouts on that scale are signal-controlled anyway these days (at least here in the UK) so a separate system of crossings to allow cyclists to move around the outside like pedestrians can be provided.
> The risks you describe are all true, but mostly apply at traffic lights as well: a cyclist still has to navigate into the right-hand lane at a junction if turning right from a two-lane (or more) entry road.
No, they don't - as a cyclist, you can do a hook turn: http://www.cyclingtipsblog.com/2009/05/the-hook-turn/. Even if this is technically illegal in your jurisdiction, I'd still do it, since it's much safer than trusting car drivers to do the right thing.
Solid islands are also dangerous, since they narrow the road and most drivers will try and push past rather than slow down to cyclist speeds. Neat fact: the handlebars on my old bike are higher than the wing mirror on a VW Golf. How do I know this? They passed me on a roundabout, and the wing mirror passed under my handlebars. Lucky I ride a bike with flat bars, and not a racer.
May I ask where you cycle? What you're describing is very different to my experience here in Cambridge, and I can honestly say that I have never seen any cyclist of any standard pull what you call a "hook turn" here.
Given that we have a vocal (to put it mildly) local pro-cycling campaign and there is another one down the road in London, while I'm not questioning your own experience, I am a little surprised that no-one has been talking about and promoting such an alternative cycling technique if it really is safer in general.
In fairness, drivers around here are also very familiar with and aware of cyclists. While some drivers pass too close, and some cyclists complain at any driver passing less than an absurd distance away, for the most part the two groups do actually get along. I've never had any problem waiting to turn right on a normal line in mid-junction, nor with drivers passing as close as you describe while going around a roundabout.
Roundabouts typically take up a bit more room. You would need to cut chamfers off (demolish and rebuild) the corners of buildings at junctions to make enough room for a roundabout that's more efficient than traffic lights.
Point of note: the average dwelling in the UK is 75 years old. That's a mean; a substantial proportion are over a century old, and many are under preservation orders ... and they were designed at a time when automobiles either did not exist, or were owned by fewer than 2% of the population. (Car ownership hit one vehicle per 25 people only some time after 1945.)
We just don't have the land area to go throwing up new construction (including buildings or roads) willy-nilly -- if the US land area was populated to the same average density as the UK there would be around 6 billion people living there, not 300 million.
This goes for most of the living urban cores of European cities, similarly. Urban sprawl is cheap and lets you build around the automobile -- although it's intrinsically hostile to public transport by reducing population density. But if land is astronomically expensive to re-purpose, urban sprawl simply isn't a viable solution to the transport problem.
Of course the average population density of the UK is a bit misleading because significant chunks of our islands are rather thinly populated - particularly the Scottish Highlands and Islands.
That means we're pretty tightly squeezed here in the bits that are densely populated.
