However, blind people have expressed serious concerns about the concept of continuous pavements, which RNIB oppose. As the RNIB says, the problem is that they “pitch us, without warning, into a shared space without knowing it.”

The London Cycling Design Standards acknowledge these concerns, saying that continuous pavements “should currently be regarded as experimental in the UK. Further development of the concept is needed, in consultation with access groups, to determine acceptable approaches, given concerns over the lack of delineation between the footway and the area accessible to vehicles that runs over the entry treatment. Any proposal should be subject to an Equality Impact Assessment.” (P43)

So would the obvious design solution be to add tactile paving to continuous paving, to warn a visually impaired pedestrian that they are entering a space where they may encounter a vehicle? My hunch is that many advocates of continuous pavements would resist the use of tactile paving in such situations on the grounds that it would signal, to both pedestrian and driver, that walking priority is limited. Rather than walking across the side road with full confidence, a pedestrian should pause and check that vehicles aren’t turning into or out of the side road, undermining the ‘people over vehicles’ philosophy. Tactile paving also can inhibit the mobility of other disabled pedestrians, so it should not be over-used.​

But on many ordinary pavements, there is already a risk of encountering a vehicle every few yards - where there are driveways or garages. The reality is that many pavements are already ‘shared space’ to some degree, as is every crossing of a side road. Would anyone expect tactile paving to be installed every time a driveway crossed a pavement? (Saughton Road North, Edinburgh).

So, should continuous pavements have tactile paving? The answer is probably “it depends”. As with so many other aspects of street design, each site really needs to be designed specific to the context, especially taking account of traffic (volume and speed).

As the DfT is at last reviewing the woefully out of date guidance on ‘Inclusive Mobility’ and tactile paving it is timely to give this some serious thought. It would be useful to have detailed national guidance, based on both research and consultation on if, where and when tactile paving is required on a continuous pavement.

So the *total* number of injuries to pedestrians caused by collisions with cyclists is considerably higher than the ‘1%’ figure often quoted (for fatalities). I would imagine by a factor of at least 3; possibly as high as 5 or more?  Of course, this is still very low compared to crashes involving motor vehicles, which are overwhelmingly responsible for pedestrian (and cyclist) casualties. So does it matter if cycling accounts for 1%, 3% or 5% of pedestrian collisions? Yes, I think it does. 
​
Firstly, selective cherry-picking of statistics discredits the basically sound argument that the  principal focus of road safety efforts should be on motor vehicles (rather than cyclists).

Secondly, we need to recognise the anxiety that cycling can cause pedestrians. While there seems to be little statistical or academic evidence on this, there can be no question that inconsiderate cycling - especially on pavements, or through pedestrian crossings - seriously intimidates many pedestrians; especially those who are blind, deaf, old or generally less confident and steady on their feet. Indeed, Peter Walker acknowledges in his video that pavement cycling can intimidate; but in the next breath describes it as “harmless”. 
​
While it is wrong for the media and politicians to disproportionately focus on cyclists’ involvement in pedestrian collisions, it is also wrong for cycling advocates to simply dismiss the impact of inconsiderate cycling as negligible or irrelevant. A little more empathy from some cycling campaigners - such as on the lines of Suzanne Forup’s excellent recent blog  - would not go amiss; it might also do more to win wider public support for more investment in cycling.