Continuous footway, side-road crossings, simplicity and clarity, blending, getting it right, getting it wrong.
What it is which makes the Dutch ‘continuous footway’ design work so successfully?
Copies of this are becoming more common in the UK, but are we getting our designs right?
What Dutch design principles are relevant here? What can we learn from these principles?
What’s wrong with the idea of ‘blending’ designs?
What is ‘continuous footway’?
In Dutch cities and towns one of the features which first stands out to a UK audience (and those from most other countries) is the use of ‘continuous footway’. I’ve written about the significance of this feature before.
What I’m talking about is the situation where a roadside footway (what we call ‘the pavement’ in the UK) carries on alongside a bigger road, crossing the end of a smaller side road without an interruption. You can see this drawn in the image below – indicated with a red arrow.
There are different names used for this design – but in the UK it’s quite common to call it ‘continuous footway’. Often – as in the drawing – there is also a cycleway, but that’s not always the case so ‘continuous footway’ seems the simplest description. I’ve also heard the phrase ‘blended crossing’ – the reasons I dislike this term so much will become clear later. And I’ve heard ‘Copenhagen style crossing’ – although strangely this phrase was being used for something I never found in Copenhagen.
I should say that many Dutch people will look at you strangely if you try to talk to them about ‘continuous footway’. What we can see in the images I show here just represents ordinary Dutch street design. Continuing a footway across a road end is so normal that they don’t see a design feature here at all – they can’t understand how different this is to how most of the rest of the world works.
This is a photo of continuous footway in Amsterdam:
I was standing on the side street to take the photo – people are allowed to drive over the footway (toward where I was standing), leaving the larger road (which you can see at the back of the picture) to join this residential side street. Here are some images from a virtual model of continuous footway:
It is very important to understand that continuous footway is used in very specific circumstances in the Netherlands.
This is used as much as a ‘gateway’ – indicating to people driving that they are entering (or leaving) the access streets belonging to a residential area. At this point you need to drive up and over a section of footway (and possibly cycle track). The speed limit changes at this point. The road materials will change. The road width will change. Most importantly, the streets beyond this will be designed so that they aren’t any use to you unless you are visiting a property in the area.
You can see the sign in the image indicating that this is the boundary of a 30kph zone. This is just what people expect to encounter when entering or leaving such a zone, in a car or other motor vehicle. I’ve written much more about the residential local access streets which like beyond such a gateway in the article ‘I want my street to look like this‘ (which includes animations showing continuous footway in context).
Of course the UK does not work like this. I’d like to remind everyone how big a difference there is between Dutch and UK road design. The image below is a quick and rough impression of what this road end would look like in the UK:
The contrast (I’ve animated the transition below) is startling:
Here are another three images of the location above – from different angles (click/tap to see full images).
Here’s a quick edit to another photo – taking an image of a different location, where the side road is actually really difficult to see – and providing a very rough impression of what we’d expect in the UK:
Animating the change:
How can people think that the Netherlands is unfriendly for walking when you see that contrast? I know which street I’d prefer to be walking on.
Simplicity and clarity
Dutch streets and roads are designed with a set of important principles in mind – based on their ‘sustainable safety’ approach. I think it is essential that we understand how important this approach is in defining the way that Dutch streets work. For an introduction read Mark Wagenbuur’s article here (but you can finish reading my article first).
One of the important results of this sustainable safety approach is that Dutch design is simple and clear – in my words I think of ‘simplicity and clarity’ as, in effect, one of the most important Dutch design principles. I believe that the phrase “Zelfverklarend wegbeeld” – meaning “self-explaining road” – is used to explain that the design characteristics of a street or road should make it completely obvious what behaviour is appropriate, safe, and reasonable.
Look at how simple the street layout is here – but also at how clear its functionality is. There is no confusion – no mess – no vagueness. No special symbols or signs are required. It is not necessary to add lines on the ground. Nothing special tells people to give way. The basic design features make it completely clear that this is a space where pedestrians and people cycling have priority. It is relatively clear that you can drive over the footway and cycleway – but it is completely clear that this is what you are doing. You’re driving over the footway. This is not a place where the roadway and footway are mixed together or ‘blended’. This is a piece of footway/cycleway which you can drive over to get from one piece of road to another.
What makes the design so simple and clear?
It is fairly obvious why this design is so simple and clear – but I’ve seen such terrible copies of this in the UK that I feel the need to actually write the following list of factors:
- The surface material used on the footway continues unbroken across the space AND onward (well beyond this local feature). There are no significant changes in colour, design, or material. There are no materials or structures indicating the edge of the space to be used by vehicles, nor suggesting that there is anything special about this piece of footway. This is not an accident. The Dutch designers didn’t forget to do something here. It is an essential and inherent part of the design – making it safer.
- The footway does not change height compared to the road. It is necessary to drive up and over the footway – making the presence of the footway really obvious.
- The edge of the main carriageway continues in a straight line without any indentations at all. There is no kerb or kerb-like infrastructure highlighting the line to be taken by vehicles. There is no short section of side road allowing people to leave the main flow of traffic. This isn’t a flaw in the design. We won’t make a better job of this by changing this feature of the design. It’s a planned part of what makes the design work well.
- The cycleway also continues unbroken, with no changes in surface or material or colour or height, and with no special symbols.
- The end of the side road – where it meets the footway – is also really obvious. The road end narrows visibly so that it is only the width of one car where it meets the footway (and no parked vehicles obstruct vision).
And the ramp matters too
Aside from the visual simplicity and clarity here, a further design feature (necessary to keep the footway level) is a sharp ramp from the road surface, to the height of the footway. The footway is flat – the surface height change for vehicles happens over a very short distance – using special Dutch kerbs “inritbanden”.
I’d translate ‘inritbanden’ as ‘entrance kerbs’ – search the internet for inritbanden and you’ll get plenty of images – and this commercial sales page has nice drawings. While we’re doing Dutch words – search for “uitritconstructie” – which I’d translate as ‘exit construction’ – for images of this whole arrangement… again emphasising that to Dutch people what we’re looking at here is seen primarily as a gateway/entrance/exit design. You might also want to research Danish methods – some words/phrases to search for to get started are “gennemgående fortov” “overkørsler” and “gennemførte fortov” and this document is helpful too.
The inritbanden can be seen in the other photos – but the image below is helpful.
The ramp on these entrance/exit kerbs is sharp enough to mean that if you drove an ordinary car over them too quickly you’d risk causing damage to the vehicle.
Designs like this are not just ‘self-explaining’ – they force people to behave appropriately too.
So to summarise – a good design of a continuous footway (with or without continuous cycleway) involves:
- a simple and clear design which makes completely clear that this is an area of footway (over which it is permitted to drive when nobody is walking/cycling there);
- absolutely nothing suggesting that this is anything other than footway, no markings, no lines, no changes in surface;
- sharp ramps onto and off the footway meaning that those driving over the footway must drive very slowly.
More Dutch continuous footway
Not all continuous footway in the Netherlands is quite as good as this example – but examples this good are very easy to find.
Here are images of five more locations. Some of these locations include a continuous cycleway (off the road) but some don’t. All show the features I’ve listed above.
How these designs talk to us
Moving on from the pretty pictures – I think that it is worth looking even more deeply at how these designs work. Specifically I think that it is worth thinking about how we (as road users) “read” the infrastructure here. We should look at what signals or messages these designs send us – or if you like, how these designs “talk” to us.
The reason that it’s so important to consider this – in such depth – will become clear later, when I show the results of bad design.
Let’s take the first example again – I think that the infrastructure says these things to us (written on the image):
We need to remember that users of roads and streets are imperfect. They are receiving lots of signals – lots of messages – from the world around them.
“Here’s the road.” “This is the speed you should be driving.” “Look out, there’s a bump ahead.” “Look out, there’s a person ahead.”
And they are usually thinking many thoughts simultaneously.
“Did I lock my door?” “Am I late?” “How am I going to explain to the children?” “Why me?” “Why did he say that thing to me?” “I like this music.”
We might have special symbols or signs to communicate messages to people, but these are human beings not computers. Human beings filter the environment for the messages they most need – and designs like this communicate these important messages really clearly (above the noise if you like).
Think about last time you drove along a road you didn’t know. Did you really read every road sign? Did you read every street name? Were you carefully analysing and considering all options and decisions? Or – more likely – were you partially on ‘auto-pilot’? You might have been thinking carefully about what road to take and whether you were late for an appointment – while some deeper part of your brain took care of steering the car to keep it on the road. The point is that people filter the signals and messages they receive, usually based on their expectations, ignoring much of what they see. If our design is different from the norm then we need to make sure that the most important messages win out over the less important messages – lifting the important stuff above the background noise.
Getting it right and wrong in the UK
It might seem odd to discuss about how these designs talk to us – but that’s because in the Dutch designs shown here the messages are so clear and simple. Let’s look at what can go wrong. Let’s look to the UK.
I’m going to criticise people’s work here – if you work in delivering infrastructure in the UK you may want to read this note before proceeding…
In moving on to criticising designs in the UK I want to make it clear that I appreciate the bravery and determination which will have been needed to get these ‘new’ designs in place – to break from the UK norms and to try to improve things. The pressure to ‘do things normally’ in the UK (and almost everywhere else) is enormous. Anyone prepared to stick their neck out and ‘do things better’ deserves our/my admiration. At the same time this blog isn’t going to be much use – in achieving what I set out to do in writing it – if I only talk about what works and not what doesn’t. I’m not aiming to criticise people directly involved in the designs I show here – I’m actually doing my best to show where compromise leads to bad results, and I’m pleading that the same people are in future given the support and tools they need to do the job properly.
Here are two examples from Glasgow. These are actually pretty good and work well – despite some failings when compared to the lists above. I’ve included them to show that some compromise can be allowable in some circumstances. These are on sections of street which have high pedestrian usage, and cross the end of very narrow lanes. Visually they still look a little like roadway from some angles – but the buildings and the crowd – the sharp ramp, and a fairly clear and straight carriageway edge – make them pretty effective. Interestingly I’ve walked over these for years without noticing them. Someone did a good job here – but I’m still going to argue for ‘better’ next time.
You might think of it like this (referring to the image below). The nice straight kerb edge, and the continuation of the double yellow lines, both tell us that the carriageway is the black stuff. There is also a message from the change of materials on the footway – suggesting that this might be regarded as a piece of roadway – but this message is very much weaker. I’ve indicated the relative strength of these messages with the colour and size of what’s drawn/written on the image.
Here’s a more compromised second example from Scotland. (When I first encountered this I said that I thought it was the first continuous footway in Scotland – but since noticing the Glasgow examples I’ve had to eat my words.)
The image below shows what this design looked like when I first visited. I thought that it was not perfect – but I observed that it seemed to be working and I was enthusiastic.
Here is a very short video of the effects – showing that the person driving here understood what was expected of them. To those in some countries (where it is customary to wait for those on foot) this will look unremarkable – but to those of us in the UK this behaviour is remarkable.
Now, based on my list of essential features it’s clear that this design isn’t perfect. In particular, the kerbs and cobbles suggest a road end. We receive what might be called ‘blended’ messages from the infrastructure – about how to use it, about who has priority, and about what the purposes are of each surface.
But the design is probably just about clear enough…
…except that when I visited again a few weeks later traditional road markings had been added:
With one tiny bit of paint we’ve VERY significantly strengthened the messages – for those driving, looking from the road where I took the photo – that this is a roadway. Yet many of the most important design elements, viewed from the perspective of those on foot, continue to suggest that it is footway.
The problem in more detail
This is what I mean by the blending of designs or messages. There are now mixed messages being given – both to those on foot and to those driving – messages making it unclear whether this is road or footway.
Which is VERY different to the Dutch design – where it’s obvious TO ALL USERS that the footway is footway – that the cycleway is cycleway.
This is a problem. People driving – accustomed to seeing roadway – driving in a manner based on their expectations of normal UK behaviour – are likely to see the roadway first and other features only as secondary (if they notice these at all). But those walking or cycling are relatively likely to see footway/cycleway first, and notice the roadway second.
I’ll draw lines again on the images to emphasise my point…
Faced with a complicated and confusing situation – people driving, LOOKING FOR ROADWAY, will see first a roadway. They’ll mostly take notice of the stuff I’ve highlighted with red lines – taking only secondary notice of the stuff I’ve highlighted in blue.
Why don’t they see the stuff marked in blue? Well it’s because people driving are receiving lots and lots of messages, all the time – they’re not just seeing and receiving messages from the infrastructure, but also the people, the shop, the sign, the buildings, and from lots more sources. Rather than the simple messages in the Dutch design – which are strong enough to carry above the complexity – to be heard above the noise – here we’ve blended everything together. So someone driving is receiving messages like this (full size image here)…
…which is hugely confusing.
We’re used to dealing with complex and confusing situations when we drive – we deal with them by filtering out the confusion, looking for the main features we expect to see…
Of course we CAN see that something is different here. Most people, standing where I took the photo, with time to spare, will see that something is different. But these messages that something is different have to fight for attention above the noise and above our expectations about how roads work here in the UK. And these messages need to be processed in a second or two – as someone approaches this junction at 30mph.
Which is problematic when those on foot and cycling see what I’ve outlined in blue before they see what’s outlined in red:
Blending is (usually) bad
Somewhere – somehow – a strange idea seems to have emerged in the UK. The idea is that blending normal designs together will give us something safer. The idea is that people, driving, cycling, and on foot, when faced with a confusing design, will slow down, take care, look out for others. The simple and clear Dutch continuous footway is seen as a more dangerous design – and mixed up and compromised designs are seen as safer.
Now to be absolutely clear – it certainly seems to me that it is possible, through design, to change how quickly people drive, or how much attention they pay to others. Making someone who is driving less certain about the priority they have over those on foot might make them slow down a little. That can be a good thing. There are even situations where AFTER vehicles have been made to travel more slowly it can then also be helpful to increase the confidence of those on foot that people driving will wait for them. But designing consciously in this way isn’t the same thing as just blending designs together so everyone is faced with more complexity. In many places the blending of the messages that we receive from infrastructure simply makes things more dangerous.
Here’s what I might describe as a useful rule of thumb:
If you can design so that you clearly signal to road users that something works differently at a particular location, you might cause them to take extra care. But if instead you design something that people find confusing, they will filter out the confusing signals and ignore them, and will look for what’s normal, and will behave as normally as possible using only what they can see as normal as a guide.
One of these results makes people safer. One of them puts people at additional risk. The difference between a design which creates the first result and one which creates the second result is very very fine.
If we’re going to introduce infrastructure like Dutch designers use to make their towns and cities more people-friendly (and I hope we are) we almost certainly need to do it properly. The Dutch designs aren’t the way they are – safe, usable, and understandable – because people there are somehow trained to behave properly and to understand them. The Dutch designs are the way they are because this has proven to be the best way to design them – this has proven to be the best way to create the desired effects in all users.
The Dutch designs are designed the way that they are because this is the safe way to design them.
How can we have found ourselves half-copying a design and assuming somehow that half-doing something properly makes it a safer design? This is not true. It does not work.
Blending can be very bad
As an aside from the consideration of continuous footway, another very different Scottish example provides a really clear illustration of why ‘blending’ designs is (usually) a bad idea.
This was pointed out to me as being a site where there had been issues between people driving and people walking. When you look at the design that’s not at all surprising. This is what people see when they are walking up the pedestrianised street.
What’s not obvious is that a road – carrying large and sometimes fast vehicles – crosses the street here.
Observe the behaviour of the truck driver in the video. He/she doesn’t slow down significantly – meaning that at least one of those crossing is put at some risk (had someone stumbled or fallen the truck would not have had time to stop).
Given normal UK road behaviours this isn’t surprising. To someone driving this looks like a normal road (images looking at the crossing from both directions):
So to people who are driving this appears to be almost a normal road (although to be fair there are some restrictions on what vehicles can use it). But from the pedestrian street it appears as a continuation of the pedestrian street. How is that a good idea? I like the idea of removing the certainty, for those driving, that they have priority… but that’s not what happens here.
If a design is to have an effect on safety and priority then the visual effect of the design FOR THOSE DRIVING has to make it clear that people on foot will be taking priority. It might be half-hearted in indicating to people on foot (or cycling) that something is different – but it can’t be half-hearted in indicating to those driving that something is very different. Remember, this is the UK. It is not unknown for people to use vehicles to drive AT a pedestrian who is in the way – seeking to frighten that person so they learn not to do this again. Mild visual uncertainty will simply be ignored – and people (driving) are very likely to assume that the road continues in the normal way – and that they should behave as normal – not even slowing down.
There are relatively simple ways to fix this particular design I think. Probably the road needs to narrow considerably. It’s a two way road (surprisingly). Forcing this into one lane would mean people having to slow to wait for one another if there was oncoming traffic. This would increase safety in itself. The defined crossing area needs to have a much stronger visual presence – and it needs to be VERY MUCH wider. A sharp ramp would physically prevent people from driving fast. And something is needed to break up the view of a roadway (wide, with straight kerbs) going into the distance if a driver looks further ahead.
Better still – if we want a pedestrianised street (and well done those who made this happen by the way) maybe we should design a pedestrianised street which you can’t drive across at all.
Other side road crossings
Sometimes, rather than attempting to create continuous footway, Dutch engineers use an alternative approach. Here are a two examples. I include these not because I like the design very much, but because the design remains clear. The standard red tarmac screams out the message ‘cycle track’, and the standard ‘elephant footprint’ white markings also scream out ‘cycle track’.
Both designs are compromised – but both remain relatively clear.
When people outside the Netherlands compromise in similar situations then the results can be really pretty poor.
I don’t think this one works at all – not even slightly… it makes the ‘blending’ mistake again.
How much attention do you think those driving vehicles in this confusing environment have for noticing the green stripe? Compare this design against my checklist above. Consider what I have written regarding messages/ clarity/ simplicity. The issues here become obvious.
I’ve picked on this one example, but others are easy to find. Here are two more. Neither are good:
In these examples, from the perspective of someone driving, the message that there is a crossing here – which it is utterly essential to communicate – is lost in the confusion.
One hope I have is that at some point somebody will begin to work on a standard UK road marking suitable for indicating the presence of a cycleway crossing a street like this. Elephant footprint markings might be ideal as part of the solution – where a continuous footway solution isn’t suitable – IF they had some matching rules to support their safe use. They are now allowed for indicating crossings at signals. And surely it is time for the UK to agree on a colour to be used to indicate infrastructure for cycling. By far the most powerful indicator of a cycleway in the Netherlands is the standard red/brown colour (not used for anything else).
UPDATE – I’ve written more on this issue here: Imagining better designs…
Here in the UK – we’re so busy protecting our standardised systems and symbols from any change at all that infrastructure to support cycling (which doesn’t fit this limited system) is being delivered completely differently in different locations – varying even from street to street within one city. In the Netherlands, where they are good at experimenting and trying new ideas, and where I’m told ‘pragmatic rule breaking’ is a national habit, they’ve ended up – as a result – with well standardised systems.
So to sum up…
Good looks like this:
When we are trying to change how people behave on our roads, half-heartedness multiplied by half-heartedness often produces something actively dangerous, not something safer, and not even something half safe.
A half-hearted attempt to signal pedestrian or bicycle priority can create something more dangerous than if you hadn’t bothered at all. A half-hearted narrow cycle lane on a road can make the road more dangerous to cycle on than a road with no cycle lane.
With rare exceptions, blending the messages we receive from infrastructure – blending two safe designs together – doesn’t produce something safer, but something more dangerous.
If we’re to find a way to support people to re-engineer our streets, roads, towns, and cities – making them pleasant and safe to walk and cycle around – then we need to actually allow them to implement the designs which have been proven to work elsewhere. Dutch people are just people. Danish people are just people. Their designs work because they are good designs, not because the people are any different. There are differences in law and habit which we have to work with in the UK – and there are differences in the road markings and signs which people are allowed to use – so we can’t just copy everything and expect it to work. But we CAN change the law and we CAN change the road markings we have available.
Perhaps I can conclude with this idea: Making a safe Dutch design less Dutch and more ‘UK-ish’ won’t make it safer if the ‘Dutchness’, and the level of safety it provides, are one and the same thing. So next time someone tells you that they are adapting a Dutch design to make it more suitable for use in the UK, please ask them to explain exactly what they have in mind…
These are just my opinions. If you have an alternative view, or you’re Dutch and want to correct my interpretation of the infrastructure and policy in your country, or if you have good examples of continuous footway in the UK which I missed from the list, then do please comment below.
Streetview list – continuous footway, UK.
The following links to Google Streetview images provide other examples of continuous footway in the UK. These are at least attempts to do the real thing. Some of these examples show really good design. Some of these examples show really bad design (and the decision to include them as ‘continuous footway’ at all has been difficult. I’ll leave you to work out which are which. The new shiny ones aren’t always the most effective – look for the design features I’ve described before you are fooled by new materials.
Leicester: 1 | More recently the road was closed: 2
Convent Walk + Glossop Rd, Sheffield: 1
Clapham Old Town, London: 1 | 2 | 3 | 4 | 5 | 6
Cambria Road/A2217, London: 1
Stapleton Road/Upper Tooting Road, London: 1
Lower Marsh, Waterloo, London: 1 | 2 | 3 | 4
Kennington Park Road, London: 1 (see also video).
Hoe Street, Walthamstow, London: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11
Forest Road, Walthamstow, London: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12
(These Forest Road links, above, are to slightly older Streetview images – look also at the 2018 images for the same sites – what has breaking the flow of the cycle lane across the road ends done in terms of changing the visual signals about priority?)
Francis Road, Waltham Forest, London: 1 | 2 | 3 | 4
Deptford High Street, Lewisham, London: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8
(Note that I have intentionally not included some junctions further south, like the one seen in this Streetview link. I do not regard these as including ‘continuous footway’ because of the lack of some of the important defining features I’ve outlined.)
Bearsden, Glasgow: 1
Prestonfield, Milngavie, Glasgow: 1 | 2 | 3 | 4
City centre, Glasgow: 1 | 2 | 3 | 4 | 5 | 6
Castle Boulevard/Abbey Bridge – Nottingham: 1 | 2 | 3 | 4
Walton-on-the-Naze High Street (Essex): 1 | 2 | 3 | 4 | 5 | 6 | 7 (from Ranty Highwayman’s post)
Mugiemoss Road, Aberdeen: 1 | 2 | 3
Newbiggin-by-the-Sea (Northumberland) : 1 | 2 | 3 | 4 | 5 | 6 |
If I missed any examples let me know – this list only includes examples of continuous footway, not other side road crossings / side road entry treatments. If your local example isn’t in the list it may be simply because I don’t think it’s actually an example – even a bad one – of continuous footway at all. There are lots of these around. I greatly dislike them because they fail against the main point made in this article… which is that clarity is important and that ambiguous design is (often) unhelpful or dangerous.
Since writing, I note that the Dutch ASVV ‘Recommendations for traffic provisions in built up areas’ (1998) specifically gives instructions as follows (heavily edited from several pages): “Uncertainty about the right of way situation can be considerably reduced by having a clear road lay-out, where the subordinate importance of the adjoining road is accentuated… In order for a construction to make clear what the status of an exit is… there should be a clear difference in the hierarchy of the roads involved. Exit constructions should accentuate the difference in character of the area beyond them… In other situations no exit constructions should be applied, but alternatives should be sought (for example by regulating right of way with traffic signs).” Also: “Exits where there is no difference in elevation should be advised against.” And later: “Footpath and possibly cycle track should be continued as far as possible in terms of colour and structure… supplementary elements such as road marking, traffic signs, or other road marking are all possible but not recommended.”
- I want my street to be like this (detailed discussion of Dutch residential local access streets – the areas which are behind the gateway created by continuous footway)
- Imagining better designs… (looking at options and a way forward where continuous footway isn’t suitable)
- Amsterdam vs Copenhagen part 2 (which includes Streetview links to views of Netherlands junctions)
- What nobody told me (introduction to Netherlands urban design)
- Read everything – New here? This is a suggested reading order.
- Visual Priority (excellent article on the same thing from the Alternative Department for Transport)
- Continuous paths across minor junctions from the same source.
- A view from the Drawing Board (Cycling Embassy of Great Britain)
- We can rethink the little things to improve walking and cycling (Ranty Highwayman’s 2013 post begins with discussing this kind of arrangement – we may also have him to blame/thank for the UK phrase ‘continuous footway’ – see tweet)
- Driver behaviour at continuous footways research (Steer Davies Gleave, March 2018). Report produced for Transport for London.
- Nader onderzoek uitritconstructies en voorrangskruisingen (“Further investigation into exit constructions and major junctions”, SWOV, 1998) – you may need to use a translation service to access sections of this text individually
- Scroll below for comments.
If you landed here from a Twitter link on a mobile device you may need to press ‘Leave a comment’ below to see the comments on this article or you can reload the page to see them.