There's a busy thread about segregated infrastructure going on at the excellent 'ibikelondon' blog.
Here's my spin on things. First of all,we support segregation at Cycalogical. Fear of traffic is the main factor stopping most people cycling, it is difficult to see how that can be addressed without segregation.
1. Martin Luther King said "I have a dream". He didn't say "It's a great success that black folks can get on the bus, even though we have to go to the back and give up our seats when white people want them". Now I'm not drawing any close parallels between cycling and the civil rights struggle. Cycling is purely voluntary, and as soon as we get to work and put on civvies we cease being members of an ill-treated minority. My point is simply that while you have to accept the world is the way it is, you also cannot allow people to delude themselves into thinking that it does not need to be better or that it cannot be better. Having ideals does not make you naive.
2. Most of the roadspace in the UK, particularly in London, is unused most of the time. It is only the bottlenecks that are busy. Having two lanes of traffic does not necessarily make the traffic move faster, because the bottleneck (usually a roundabout or light-controlled junction) is the limiting factor. Once the capacity of the bottleneck is reached, traffic will back up. Having two lanes simply means the queue is shorter, but the time spent queueing is the same. The only advantage the extra lane brings is that it will take longer for the tailback build up to the previous junction, thus delaying the onset of 'gridlock'. In other words, if you remove one traffic lane where there are two, although you halve the theoretical capacity of the road, in practice the road cannot operate at capacity. As the Design Manual for Roads and Bridges puts it, "The potential capacity of a link will not be reached if either the capacity of junctions along the link or the capacity of the adjoining network is lower than the link in question." In other words, it is a false premise that there is no spare road capacity that could be given over the segregated cycle lanes.
3. There are quite a number of wide minor roads, which could quite easily accomodate a segregated cycle lane, even without removing traffic lanes or parking spaces. Four examples: Magdalen Road, Wandsworth; The Chase, Lambeth; Prince Consort Road, Westminster; Horse Guards Road, Westminster. Once again, it is a false premise that there is no available space for segregated cycle lanes.
4. There are fairly minor interventions that can free up space. In a typical minor road, there will be parking on both sides and one or two traffic lanes. You can rearrange the road so that herringbone parking is allowed on one side, and the other side is available for a segregated cycle lane. The general traffic lane might have to be made one-way.
5. Segregation is not necessary where traffic volumes and speeds are low. Simply closing one end of a road to make it a cul-de-sac for motor traffic is often sufficient. It would be easy enough to set up a legal framework where only motors used by residents and legitimate visitors are permitted to enter a road (i.e. "no entry except for authorized vehicles"); local communities could police such arrangements themselves.
6. Conflict with Pedestrians. Most shared paths and segregated cycle paths I use, there aren't enough pedestrians to cause problems most of the time. The main issue is there are so few segregated cycle paths, and so few cyclists, pedestrians don't know how to behave around them. They'd get the hang of them quickly if there were more segregated lanes and more cyclists.
All of this is not to say that a segregated network could be set up without needing to solve some major problems. However it is to say that a network of low-traffic and segregated routes safe enough for kids to cycle to school could be set up while maintaining car ownership at current levels. Nothing particularly revolutionary is necessary In other words, we are dealing with a soluble problem. What is lacking is the political will.