Monthly Archives

9 Articles

Posted by on

The Future of Twitter

It may surprise you tech-savvy reader but most people still don’t use Twitter. Though we’re one of the most populous Twitter nations (4th in the Twitter league as of 2012), only one in five of the UK population uses Twitter from the stats I can find: i.e. users are in the minority. And of those, around 40% are passive — readers — rather than active — writers.

All this explains why yesterday around the network’s seventh anniversary I found myself on BBC Radio Scotland explaining the basics of Twitter. And why it seems like a good time to analyse the success of Twitter.

That success might in large part be attributed to the old adage that the simplest ideas are the best. Twitter took one of the key ingredients of the early social networks — the status update — and turned it into a hyper-concise broadcast medium. In tune with our time-pressured lifestyles it took what was most valuable to us about social networks — what other people were thinking, doing, eating, seeing — and condensed the ability to share and read this information down to a compact form.

Though there have been numerous evolutionary changes in the way Twitter operates this simple core has remained, and grown in popularity. Today around 200 million people use Twitter regularly. And from these 200 million users, Twitter expects to generate a billion dollars in advertising revenue this year.

The way people use Twitter has evolved too. One thing I have noticed recently is a willingness to conduct business via open @ messages that would previously have been restricted to private direct messages or SMS. For example, organising where and when to meet.

This says a lot about our attitudes to privacy in this social century, but it also comes in part I think from a willingness to jump fairly indiscriminately between multiple short messaging services. Conversations flip between Facebook, text, Whatsapp, Skype, Twitter, and more, albeit not seamlessly.

It’s the friction in swapping between these services that makes me wonder again if there isn’t an argument for a standardised, internet-based short messaging protocol. This would clearly damage Twitter and all the others unless they could find a way to differentiate. But in the long term it makes little sense that we preserve all of these little digital fiefdoms if they are fundamentally doing the same thing: carrying short text messages and linked media from one to one, or one to many. If this function is really a core part of the internet then surely its operation should be on distributed, internet principles?

It were to happen, one of the barriers to this shift would be identity: without a single authority who warrants identity (and polices abuse) on a communications network? The answer might come from the old telephone network: we pay for a company to connect our telephone number to the network, and sign up for their fair use policies in the process. Small breaches of good practice are handled by our provider, more serious ones by a central regulator or law enforcement.

In this model Twitter may retain a role as service provider, or this role may be taken up by those that provide our connectivity, much as they provide many people today with their email addresses.

Whatever happens I think Twitter’s lifespan in its current form is limited. Powerful, useful and entertaining as it has become, like any social network its hold on us is shallow. Its customer base is fickle and fast-moving. And its position in the market is under constant threat from some of the world’s most aggressive innovators.

I will continue to enjoy Twitter as a regular user. But I don’t expect to be celebrating its birthday in another seven years.

Posted by on

Future Everything: Smart Cities vs Smart Citizens

On the first day of the Future Everything conference a thesis seemed to be emerging. That the future of the city was going in one of two directions. On one hand was presented the cold, planned, infrastructure of efficiency proposed by governments and major corporations. One the other was the messy, organic, joyful improvisation of citizens. These two were presented as something of a binary choice by a number of speakers, with a pretty clear bias as to which they preferred.

I missed the first speaker of the day, while giving a presentation myself elsewhere. But I caught him — Dan Hill, CEO of Fabrica — speaking the previous night at the launch event and read his theories online here.

Dan is a very engaging speaker and though my natural inclination towards concision found his essay a little extended for the argument he was making, it is nonetheless fascinating. I’ll try and précis in many fewer words here.

Dan argues that the debate around ‘smart cities’ today is dominated by large corporations promoting a vision of efficiency. When in fact efficiency is not necessarily a desirable objective for many of the key functions of a city. Rather the corporations are using this as a drive to increase network traffic and so increase demand for their own products. By contrast the real innovation is being driven by smart citizens using social media to self-organise and drive change. This progressive thinking and acting is in dramatic contrast to our rather 19th century means of decision making. Dan believes we do need to update our infrastructure but doing so through the corporate/old government approach risks leaving us with dated infrastructure that doesn’t do what we need and locks us into long term issue much like the promotion of the car over public transport has over the last 50 years.

The afternoon debate featured Usman Haque, creator of Pachube (now Cosm), designer, artist and architect and Martijn de Waal, journalist and founder of the Dutch think tank, The Public Matters.

Usman explained how the very nature of the city makes it a ‘wicked’ problem — a definition created in the 1970s to categorise problems that for various reasons are difficult or even impossible to solve. He argued that the sheer ‘messy’ complexity of interactions that make up a city mean that a centrally-planned solution will never deliver efficiency or even any desirable outcomes. Especially if those outcomes rely on the collection of endless amounts of data for someone else to ‘do good’ with.

Martijn gave three examples of smart city development from around Seoul in South Korea including Hongdae and Songdo. Songdo is the archetype of the centrally planned, corporately-driven smart city. And as you might expect it looks pretty soulless so far. What it also is, more surprisingly, is largely devoid of smart city technology — sensors, communications etc. This is all yet to come, despite the marketing hype. Hongdae is a much more organic development from sleepy suburb into jumping urban development, with relaxed planning laws enabling rapid revitalisation and renewal. His third example was of a very engaged approach taken by Seoul city hall, where they are working with citizen groups to advance problems — for example using a social platform to work with local environmental groups to help uncover where a batch of radioactive tarmac had been accidentally laid, and track its removal and replacement.

In the keynote on day two, author Anthony Townsend asserted similar concerns to Dan Hill, while acknowledging the role that companies like IBM have played in the successful smart city experiments to date. And it was something that Townsend said that really crystallised my issue with the binary choice between soft and organic, and hard and corporate.

He suggested that while one is like the mainframe, the other is like the internet: created in a distributed fashion from the ground up. Well the internet may have been built on open standards, and many of its outposts may have been constructed by hobbyists and communities. Indeed this is where the most interesting developments are happening today. But the superstructure of the internet was built by commercial entities, investing and often losing vast sums of money deploying global fibre, data centres and networks. Without that corporate investment, who would have created the incredibly expensive inter-city and inter-continental links that now bring us our videos, mails and social networks?

I believe that smart cities will be the same. There is a level of infrastructure required — power, waste, connectivity, sensing — that almost certainly will not be built organically by community groups. There will be outposts of DIY but mass coverage — the superstructure — will need to come from commercial entities, some of which will succeed, some of which will fail. We will have to trade something to these entities in order to make it worth their while to invest. And yes, we may not like those costs in the long run. But ultimately they will drive progress faster than if we leave everything to hackers and organic growth.

The analogy I offer is the difference between a house and a home. You can build your own house, and some people do. But for the vast majority, we buy a house. We don’t like the long term cost and the house we choose will usually be a compromise. But we turn it into a home with the changes we make: the sounds, smells, and colours that we add.

Cities are the same: it’s going to be a long time until society evolves past the point where the market is the primary driver for investment. And until that point if we want our cities to be smarter, we need to swallow a degree of corporate investment and recognise that the humanity, the interest, the things that make great cities, are the things that we build on top of this superstructure, not in place of it.

Posted by on

Sammy the Boblebot Part 7: Testing Times

There’s a good reason I talk and write about technology rather than designing it like my university peers (some of whom are now in very prestigious positions). I could blame the fact that I succumbed to the triple temptations of wine, women and politics when I should have been studying, but in reality I’ve always lacked some of the rigour required to be a good engineer. For example, I love making stuff and hacking around with software, but when it comes to testing and refining, I tend to get a little bored.

This may explain why it has taken me a couple of weeks to get to the next stage with Sammy: testing the sensors and calibrating the motion against their output. Though get to it I have, and the results have been very informative.

So far I have only tested the mouse sensor in the Y (forward and reverse) axis, and the motors. Here’s how I went about it.

Method

I wanted to test three things:
– What was the sensitivity of the mouse sensor in the X and Y axes
– How far from the floor would the mouse sensor continue to operate
– How did the units on the sensor relate to the time/speed of the motors and actual movement

To do this I laid out a strip of lining paper on the flat floor of the Man Cave so that I could plot measurements and provide a nice consistent surface for the sensor. It also gave me some straight lines to track movement against — easy to see if Sammy is straying off to one side or the other.

I wrote a new loop for the programme, moving Sammy forward at a stately rate with a variable interval between movement and polling the mouse sensor. Varying the delay allowed me to dial in a sensible ratio of movement to polling.

Here’s what I learned:

Mice Are Short Range Animals

The first thing I learned is that the mouse sensor only works when it is pressed right up to the floor. Even a few millimetres separation and its output didn’t just degrade: it disappeared. All or nothing. Given that I was relying on the sensor’s output to navigate and make course corrections (below) this is problematic, however…

The Mouse Sensor is Awesome

When it works, the mouse sensor is fantastic, giving very consistent readings for movements in X and Y. This shouldn’t be a major surprise: we all rely on mice on our PCs every day and this one has barely been modified. But it’s pleasing to know that if I can overcome the range issue this will be a really useful piece of kit. However…

Initialising Issues

From the start I have been having problems initialising the mouse sensor when the Arduino is first powered up. I had put this down to the dodgy hookup wiring but now that everything is soldered on or connected with proper plugs/sockets I can’t blame that any more. There’s clearly an issue. Once it’s on, it works fine, but getting to that stage requires cycling the power to the mouse sensor during the start-up sequence. This is not ideal, especially since the only way to cycle the power at the moment is by pulling off the grounding pin for the mouse.

Any Which Way But Forward

The current chassis and geared motors are really bad at moving Sammy in a straight line. Even on a flat surface, uneven gearing and the very wobbly castor mean that straight line motion is going to be hard to achieve without good course correction.

50ms@128 = 75px = 2.5mm

Starting the motors at Normal speed (128 on the PWM outputs of the Arduino) and running them for a delay of 50 milliseconds before bringing them to a stop gave me a change of roughly 75 pixels on the mouse sensor and 2.5mm in movement. These numbers are approximate: I haven’t done a proper average on the pixel count yet but the actual average (mean) movement across 25 tests averaged to 2.369mm. Understand the pointlessness of these decimals: the measurements were taken using sight measurement of the position of a 0.5mm thick metal strip against a ruler on which I could — at best — estimate position to +/- 0.5mm. I’m sticking with 2.5mm because that was the mode result and frankly because it is good enough for now. See, told you I lacked rigour…

Next Steps

I’m going to persist with the mouse sensor as it is for now: I can rely on it sufficiently for testing while it is on the nice flat floor of the Man Cave. I will see if I can get another one of my mice working with the Arduino though in order to have a piece of hardware to test with that isn’t part of the main body.

I’m going to experiment with moving the mouse initialisation around in the code to see if I can overcome the issue without having to create some new hardware to cycle the power. I’ll also try playing with the initialisation sequence — maybe it needs a delay between the reset and mode commands? Worst case is a hardware fix: It will be annoying if I need to use another pin just to reset the mouse sensor.

Finally I need to do some work on the chassis. I’m loathe to start work on the next stage of the base while I still have lots of testing to do, but I may have to. First though I will see what I can do to make this one more reliable and stable

Beyond this the issues become code again: I need to swap out my home-brew course correction for one using a PID controller, something that has been dredged from the depths of my engineering degree memory by browsing around robot control forums.

Then we’re into dream territory: I’ve been looking at ROS and considering how nice it would be if I could remotely program the Arduino by dialling into Sammy via a local host computer over WiFi. Hello Raspberry Pi. Did I mention it’s my birthday soon?

Posted by on

Sammy the Boblebot Part 6: Body Building!

As noted here, I felt it time to give Sammy a little more humanity. Or at least make him look a little more fun and sci-fi.

This isn’t an entirely frivolous exercise: there are a few reasons why now was the right time for a little hardware work. For a start, I was running out of space a little on the simple base. I was also getting some reliability issues with all the hookup wire running via the breadboard, so I figured I’d take the chance to replace as much as possible with proper plug connectors and solder/Veroboard where required. Finally I wanted to start to enclose some of the primary electronics to keep them away from little hands and the things that they tend to spill/smear.

At some point I expect to replace Sammy’s wheeled base with something a little larger and more stable, and maybe ultimately with some legs. So I wanted the new body to have some longevity. That meant being reasonably large, open, and easy to bolt on to. Knock-off Meccano to the rescue again! A crane kit I bought ages back with this project in mind had lots of lovely long right angles that could be bolted together to form an open box.

Then it was a dive into the treasure chests that are my hoarding boxes to find a suitable enclosure for the electronics. I came up with an old shaver case that clips closed and was just the right size for the Arduino and breadboard (gratefully received from my CANDDi colleague and fellow geek Oli Woodsome time back). I’ve kept the breadboard in there so that I can prototype and test new stuff before having to make it up on a ‘proper’ board. At the bottom of the ‘backpack’ I’ve kept a little compartment that can be used for stuff — though what I don’t yet know.

For the head I have an old spherical clock, the stem of which was used on my Wi-Fi antenna project a few years back. This will be filled/decorated at some point and given motion, but for now it looks suitably sci-fi.

To complete the effect (and if I’m honest, in an attempt to return a little balance to the now wheelie-prone Sammy) I added a couple of simple arms. I’ll need to do more to return stability once the rebuild/rewire is complete.

The first proper piece of rewiring has been to solder the PS/2 port for the mouse down to a piece of Veroboard and add some plug connectors. I’m hoping this will iron out some of the instability I was getting with the mouse not initialising properly until the ground wire was wiggled. Only took me two attempts to get the pinout right this time around, and fortunately mixing up the clock and ground lines doesn’t seem to have done the mouse any harm.

Next I need to wire up a proper plug into the H-bridge motor driver, though I’m tempted to keep the LEDs I’ve had in-line as these have been very useful for debugging. More Veroboard has been ordered (I’ve been using up what I think was an offcut from my last robot project, probably from around 1990).

Also still to do is sorting out the connections on the side of the Arduino board that is now facing down. These are now obstructed by the bottom of the backpack. Though the board hinges for access to the USB and power ports, it would be nice not to have to bodge it like this. Right-angled connectors running out to a breakout panel with USB connection, power and reset switches would be cool.

But first I need to bring it all together and get Sammy doing something vaguely robot-like…

Posted by on

Sammy the Boblebot Part 5: Servo Sweep

Quick update to Sammy’s progress: the servo I installed in part 4 is now functional and sweeping back and forth. I haven’t written any new functions for this yet, though the supporting Servo library probably gives me everything I need.

The new loop will probably look something like this:

  • Go forward until you get near an obstacle
  • Look left, look right
  • Turn in the direction with the least obstruction
  • Go forward until you hit an obstacle

Not very sophisticated but one more component that can ultimately be combined in making Sammy do some smarter stuff.

Posted by on

Microsoft’s Fine: What Does it Change?

£484m is a sizeable chunk of change for any company to find. Even one the scale of Microsoft, for whom it represents only 1% of revenues. This is the fine that the EU has just levied on Microsoft for breaching the agreement the company made when initially found guilty of abusing its position of market privilege.

The breach itself is bizarre. It involves the disappearance of a screen that suggested alternative browsers (other than Microsoft’s Internet Explorer) to people installing Windows or using it for the first time after puchasing a PC. Microsoft has explained that an engineer left out the line that told this screen to appear when publishing an update to the Windows 7 operating system. This is just about plausible, though it would require an awful lot of people to miss this pretty important exclusion.

What is less plausible is that the first time anyone at Microsoft noticed was 14 months later. Think of all the times Microsoft employees and its resellers must have installed or used a new copy of Windows in that time. Is it really possible that none of them noticed the missing screen — part of a legally-binding settlement negotiated with international governments? Or that they did notice but it took Microsoft, one of the world’s largest software companies, 14 months to implement a solution?

This is the way things happened says Microsoft. And hence the fine is the size it is: the EU could have increased it tenfold. Good thing for Steve Ballmer, Microsoft CEO, they didn’t: the apparent mistake has already cost him part of his bonus.

Even if it isn’t the size it could have been, this is still a sizeable fine. But will it change anything? Will Microsoft, or for that matter other global technology companies, think twice before pushing the boundaries of what is acceptable in corporate behaviour?

I think probably not.

Turn the situation around. Imagine that Microsoft, rather than its infamous ‘embrace, extend and extinguish’ strategy had decided to play nice with its upstart competitors. Not just the browser companies that were the subject of this case but the other innovators that Microsoft has crushed since the advent of the consumer Internet. Like RealNetworks for example, a somewhat unloved company due to the increasingly bloated and salesy nature of its product. But one that did much to popularise internet video and audio with consumers*.

Imagine Microsoft had stayed out of the Internet apps market and actively promoted these company’s to its users as examples of the great software available on the Windows platform. What would shareholders have made of that, especially once applications like Google Docs moved into the browser, competing directly with other Microsoft core products?

Answer? They wouldn’t accept it. Shareholders have a right to expect that the companies in which they invest strive to grow and do well. The very nature of a company of scale like Microsoft is that it has to press home the advantages it has. And its position is its biggest advantage, especially in the face of small, innovative competition. Because history has shown that beyond a certain size, companies struggle to innovate.
That after all is the story of Microsoft’s explosive growth. It was the computing behemoth IBM’s inability to innovate when faced with competition from early Apples and the new generation of cheap personal computers that gave Microsoft its first big leg up.

When assembling the hardware and software for the original PC, an IBM skunkworks team was assigned to work outside the rigid corporate structures that were slowing product development. Rather than face the bureaucracy and politicking required to get an operating system for the new machine developed internally, the PC team licensed MS-DOS. The rest, as they say, is history. Microsoft pulled in a tasty licence fee from each PC sold. And many more were sold than IBM expected — not just by IBM but by small hungry competitors who found they had access to the same components that the IBM skunkworks team had selected. By the time IBM launched its own OS, Microsoft was flying.

I’m not saying there isn’t a third way between embracing competitors and trying to smother them. But it is an incredibly hard course to plot. While a regulator may hit you with a fine every few years for bad behaviour, nothing will eat away at a company day in, day out, like competition you know you could — and for the sake of your shareholders maybe even should — squash.

For that reason I think we will continue to see companies pushing the boundaries of acceptable corporate behaviour. And every now and again, regulators vainly trying to reign them in.

*Just a note of interest, I worked on the UK PR account for RealNetworks from the end of 2000 through to about 2003/4 and spent some interesting times with the company’s executives including Joanna Shields, subsequently of Facebook and now CEO and chair of Tech City.

____

Open vs Closed: Standards for Success

In the futuristic Charles Stross novel Glasshouse, the protagonist acknowledges a turning point in human history when data was opened up not locked

Posted by on

Sammy the Boblebot Part 4: PING! Ultrasonic Bat Vision

Sammy’s movement has so far been restricted only by three things. How long I have told him to travel for, course corrections from his (unreliable) position sensor, and bumping into furniture. It is this third thing that I want to stop.

The solution is an ultrasonic distance sensor. Sounds complex but it cost just a couple of pounds from eBay and requires just a simple 4-wire hookup to the Arduino that is at Sammy’s core. Two wires provide power, one triggers the ultrasonic ‘ping’ and the other returns data when the echo is heard. Calculating back from the time between ping and echo, using the freely available NewPING library, gives you a distance between sensor and object. Awesome.

You’ll note from the picture that Sammy’s sensor is mounted on the front of his chassis, on top of a servo. The servo is ancient, reclaimed from an old remote control car. The idea is to be able to rotate the sensor through 180 degrees to enable Sammy to make ‘intelligent’ decisions about navigation and maybe even map obstacles. More of that in future.

The cable connected to the sensor is an old computer cable, of which I have many. These are now being hacked up to provide neater pin connections back to my Arduino via a Sensor Shield v4.0 — much more reliable than hookup wire going everywhere.

After a little hacking around with the code I decided to write a dedicated proximity function that could be called periodically in movement to check whether there was anything in the way, and halt movement if so. This is included in the sketch below, along with the latest version of the course correction algorithm.

RoboCode v1.4

The proximity function works fine but the course correction is still not fully functional. I need to do some concerted testing of Sammy’s movement and what the sensor is picking up before I can start tweaking the software or hardware. Though I have heard interesting reports of 35mm camera lenses being used to increase the optical mouse sensor’s range, and I have a few old cameras that could be hacked to this purpose…

More to come down the line.

Posted by Tom Cheesewright on

Should Robots Be Like People, Animals or Machines?

Should Robots Be Like People, Animals or Machines?

Why is it that we want our robots to look like us?

I’m building a robot. Having added some basic features — movement, position sensing, obstacle avoidance — I felt an overwhelming urge to make it more, well, robot-like. Or more accurately, more humanoid. And I’ve been trying to work out why.

A lifetime of science-fiction consumption has given me very strong opinions on what a robot should be: how it should look, how it should interact, and what its functions should be. Maybe I just wanted my robot to look a little more like those from films, comics and my imagination. But I’m not sure this covers it. The feeling of satisfaction from giving my robot a rudimentary head and body, even more rudimentary arms yesterday seems just a little too great for this to be the only driver.

I did some digging around the web for answers and found this fascinating paper from Brian Duffy at MediaLab Europe back in 2003. He does a very good précis of much of the research around anthropomorphism in robots, from which I can draw a few interesting things about my own desire to make Sammy (his name, given by my eldest daughter, being his first anthropomorphic characteristic) more human.

For a start humans have always liked to ascribe human-like traits to the things around us — animals, plants, computers, vehicles. And doing so in the case of robots can be particularly helpful in helping us to accept them in a social environment, and in helping us to interact with them. Adding human-like physical features has also been observed to increase the degree to which we treat the robot as if it is a being not a machine.

This makes sense: just adding a head and body to Sammy made me feel much more attached to him than when he was just a little wheeled platform. Even though this brought no additional function or intelligence.

But there are downsides to adding human-like features too. Humanoid features may not be ideally suited to the robot’s task. For example, the Roomba is a great design because it can vacuum (and live) under furniture. Giving it eyes and a mouth would just be pointless and may remove one of its primary advantages.

With humanoid features added we also begin to expect too much of what are often limited capabilities. This is an issue of which I am particularly aware. I plan for Sammy to be something of a household robot who evolves over time, in part as a means to teach my kids about electronics, mechanics and programming (though this is also a great excuse for me to indulge my geekery and create some good content for this blog). I don’t want the kids to be disappointed when Sammy can’t respond to all their spoken commands or play board games.

This leads to an interesting compromise in design: is there a sweet spot between a robot’s ‘humanness’ and ‘robotness’ that combines ease of interaction and socialisation with realistic expectations?

Pets provide an interesting analogue here. Dogs in particular. They are intelligent enough to provide useful functions (guarding the house) or fun interactions (fetching a ball), but we would never expect them to play Monopoly with us. Their physical features and voice (bark) make them easy for us to engage with and understand. But we would never mistake them for human.

So a somewhat humanoid but still clearly very ‘other’ design is probably the right approach for domestic robots of limited capability. Great for me since I doubt my restricted engineering and programming skills, and restricted budget, will allow for anything other than rough-and ready looks and limited capability.

Keep an eye on my project posts to see how Sammy develops.

Posted by on

Don’t Trust Your Instincts: Why The Future Needs Logic

Karl Neder rejected Einstein’s theory of General Relativity because it was ugly. Compared to the instinctive beauty of Newtonian physics, complex field equations and curved space time seemed fussy and contrived.

Karl Neder is a fictional character in science writer Philip Ball’s first novel, ‘The Sun and Moon Corrupted’*. But he represents a part of all of us: the part that rejects ideas and even facts because they don’t fit with our prevailing world view.

We are often so comfortable in this world view that shifting from it can be painful. I know: I’ve experienced it myself. Like Mark Lynas, author of ‘The God Species’** and environmental campaigner, I have had to radically change my views on some key issues.

Like nuclear power. I’m still not a fan of its past incarnations in either their technological form (see the waste piled up at Sellafield) or organisational (see the hidden subsidies from governments for years on end that were the only way to make it cost effective). But nuclear power increasingly seems to be our only option if we don’t want to return humankind to the stone ages and yet prevent the onrushing climate change calamity. Newer forms of nuclear power, though the capital costs remain incredible, do seem to be cleaner and safer.

Genetically engineered foods too. These days I might buy organic meat because of the welfare guarantees but I recognise that to reject GE vegetables and grains is to reject our best chance of feeding the world. Genetic engineering can increase yields, and slash requirements for pesticide and fertiliser.

While there may have been good reasons to reject nuclear power in the past, it’s irresponsible to cling to my views from two decades back when the prevailing evidence today shows that it is both safe and necessary. My objections to genetically engineered foods were based on no good sense: just instinct and propaganda (much of it created by Mark Lynas and his colleagues at the time). Accepting that I was wrong about them is vital, because people like me helped to prevent them doing years of good, feeding the hungry and cutting the impact of agriculture on the environment.

If we define our world view by one principle, rather than a set of opinions, then we can avoid making these mistakes. Or at the very least we can minimise the duration for which we are wrong and make it easier on ourselves to change our minds. That principle is a scientific one: one of decisions made on the basis of evidence.

This isn’t a principle without limits: I base my parenting on thousands of years of evolved instinct, rather consuming parenting manuals. But we don’t have good instincts for understanding the world around us in all its complexity, especially when those instincts are swayed by our desire to stay within self-defined comfort zones.

If we are to survive the future, and design a better one, it must be on the basis of evidence and logic, not instinct and fear.

*I would recommend ‘The Sun and Moon Corrupted’ to anyone, whether you are into science or not. There’s a fair amount of science in there, as you might expect from a writer on science fact, but it is wrapped around a fascinating plot, strong ideas and some really good main characters, particularly the main character and her father.

*I can highly recommend ‘The God Species’ too, whether it is reaffirming your (new) world view or challenging it. In fact I would go so far as to say it should be required reading. Maybe we need an equivalent of Gove’s scheme to give every school a bible…

Tom Cheesewright