Cycle Safety Action Plan

I think arguing about small rises and falls in the rate of casualties is probably not productive. It enables TFL's strategy of making tiny safety improvements, and then pointing to tiny improvements in safety as evidence that it's doing its best.

I agree that it isn't productive to argue about small rises and falls in the rate of casualties. This is at any rate not what Charlie was talking about. The strategy that he pursues is chiefly aimed at reducing deaths and serious injuries. We know very well that a lot of deaths are caused by HGVs. If all HGV companies adopted the safety standards adopted by Cemex following Cynthia Barlow's campaigning (her daughter was killed by a Cemex lorry), things would be much better, for instance. Reducing road danger at source (e.g., by improving the information that lorry drivers get about their vehicle and its surroundings as they drive along) is of proven effectiveness.

There is no question that TfL, owing to a lack of political direction, is for the most part pursuing business as usual, but there are exceptions, and the work that Charlie describes is among those.

Instead, we should be asking ourselves, what's an acceptable number? What should the target be? How should we define a safe and unsafe junction?
It bears repeating constantly, but junctions or roads or streets are not 'dangerous' or 'safe'. They are only inanimate objects. Even at the worst-designed junction, if people drove or cycled in accordance with road traffic legislation, there would be no casualties at all apart from the catastrophic failures, the real 'accidents', things which could probably be prevented somehow but sometimes aren't, e.g. a sudden brake failure.

For instance, in Hackney about nine years ago we succeeded in having the junction with the worst crash record for cyclists remodelled. This was a small and somewhat innocuous-looking junction at St Mark's Rise/Shacklewell Lane, where there were 10-20 reported cyclist casualties a year, including several serious ones. The problem was a sweeping left turn at a place where many rat-running motorists would turn left down St Mark's Rise and most cyclists would carry on towards the A10. The left turn arrangement was an example of bad design, but it would have been without consequences if motorists had followed proper left-turning procedures instead of putting their own sense of being in a hurry above the safety of other people. The simple fact that people often show a lack of consideration or simply can't cope with driving doesn't make junctions or any other piece of infrastructure 'safe' or 'unsafe'.

This is not just splitting hairs. It's a very important distinction to observe if you want to understand traffic and why people cycle or why not and lots of other things besides. By claiming that crashes are the (inevitable?) outcome of a 'dangerous' junction, you let drivers off the hook more than you realise. It's like blaming an uninvolved third party. You need to describe the problems with a junction neutrally, i.e. putting the finger on bad design, and separate the perpetrators' responsibility out properly.

It is still very much the case that badly-designed junctions, where some drivers' mistakes or unlawful driving cause high levels of casualties, have to be redesigned. Focusing on the design problems allows you to comprehend that problems with such junctions typically go much deeper than the relatively small symptom of bad design which are casualties.

The E&C is a good example--it is one of London's most major centres, the most important point in SE1, one of London's most important postcode areas. It has been depressed for decades following extensive war damage and the idea that most journeys should only pass through the area rather than end there. There is very little local economic activity as a result. There needs to be a complete local regeneration project, but as has been the case in London for some time, current efforts are very timid and involve little more than the Heygate Estate while putting forward only a fairly feeble suggestion for the main junction which would not have addressed what is required there.

David Hembrow's post here:

http://hembrow.blogspot.com/2011/11/most-dangerous-junctions-in-london-and.html

might be a good guide.

In what way?

In general, cyclist KSIs (per km) in the Netherlands are, I think, about a third of those in the UK. This seems an appropriate target.
Most certainly not. It is more valid to compare cyclist KSIs per km in the UK with cyclist KSIs per kilometre in NL than many other comparisons that have been tried, but it is still a fairly poor standard of comparison. Never go for just a single measure when thinking about crashes. You need multiple indexes, a mix between absolute numbers (trips, trip lengths, crashes, etc.), percentages, and qualitative information like O/D surveys, to really know what goes on and to avoid the usual problems with isolated 'targets' of having to be met, and which are ultimately usually justified with poor quality information. Basically, if it was as easy as you say, someone would already have done it.

If we set this as a target - clearly it's only obtainable through very high quality infrastructure and traffic calming on something like the Dutch model.
It depends on what you mean by those vague categories. Both 'very high quality infrastructure' and 'traffic calming' can mean a lot of different things. Don't get too heavily involved with 'the Dutch model'. Not only are there considerable differences between different Dutch towns, but alongside the good stuff there are also plenty of things that the Dutch simply haven't done very well. Yes, their transport policy history is better than that in the UK in recent decades, but there are plenty of lessons to be learned from it about how not to go about things, too.

It's also obvious that the kind of interventions TFL usually considers will come no-where close to hitting the target.
See above if you're responding to Charlie here. I certainly agree that many of the interventions TfL would consider wouldn't necessarily work.

There's also, perhaps, a more subjective safety target one could use. If it doesn't feel safe enough to cycle with a small child in a child seat, it's not yet safe enough...
Perhaps, but you can always ask for the moon. A lot places in London are places where many people wouldn't even feel safe to walk with their small child (with pollution to my understanding being one of the most serious problems for children). A more productive approach is to tackle each issue as it presents itself very specifically, and then move on to the next issue.

Real progress is usually not achieved by aiming for very remote things. If you try that, you might not see another possible step that might get you closer to a more realistic aim and which might indirectly, via any number of intermediate steps, lead to what you really want eventually. I've seen plenty of people get completely frustrated because they always had aims like the above in the back of their minds.

Sorry if that seems patronising, beside the point, or like teaching a grandmother to suck eggs, but campaigning for change is hard, long-drawn out, and sustained work.