Wednesday, 19 February 2014

The Magical Properties of Code

The magical properties of code have been much to the fore of late, notably in the car-crash launch of the Year of Code campaign. This industry lobby, to encourage us all to get coding, is fronted by Lottie Dexter, who has a background in business deregulation and rightwing think-tanks. The sight of her trying to out-stupid Jeremy Paxman on Newsnight (the unapologetically dim meets the proudly ignorant) was hilarious. Predictably, people who get po-faced about technology have slammed Dexter as an airhead PR, failing to spot that YoC is simply a PR exercise for which she is admirably qualified.

John Naughton takes a more high-minded line, rejecting the utilitarian justification that ubiquitous coding skills will help our kids find jobs: "This is first and foremost about citizenship. Today's schoolchildren will inherit a world that is largely controlled by computers and software. The choice that faces them is Program or Be Programmed". Naughton and others also noted the obvious commercial interests of the backers, e.g. Saul Klein's interest in Codecademy, and the compromised position of the BBC. More pragmatically, Jack Schofield suspects that what is on offer is little more than the basics of HTML: "how to use some parts of a mark-up language, which is (being generous) only just coding ... YearOfCode is just recycling some currently fashionable but equally fatuous American techno-romanticism".

The defenders of YoC and its ilk are perhaps more unconsciously honest about the economic reality, and thus reveal a prejudice towards techies of a certain age. According to Rory Cellan-Jones at the BBC: "There is a minority of older experienced programmers who see themselves and their craft as an exclusive band of brethren and will always be hostile to an initiative like this. A glance at the comments under a YouTube video of Lottie Dexter's Newsnight interview reveals a murky world of misogyny and coding snobbery". The gratuitous troll callout (almost every video that features a woman attracts misogynistic comments) is intended to paint Dexter as a progressive victim of reactionary forces in the mould of Mary Beard or Caroline Criado-Perez. The implication is that programming has become a 1970s-style closed shop and must be deregulated and modernised. Basically, those nerdy twats earn too much, and probably don't have girlfriends.

YoC joins a growing list of initiatives pushing the idea that everyone must "code". These range from bog-standard neoliberal lobbies to grassroots initiatives. In reality, this is a narrow spectrum of acceptably progressive positions, all of which assume that "coding" is not merely a panacea, but that it may well be on a par with religion, hence the popularity of tales about conversion and hope. The picketing of Google buses in San Francisco may be a response to growing urban inequality, but it resonates more widely as a critique of organised religion (the symbolic importance of the passengers' "separateness" and "self-absorption"), and thus the desire to democratise adept knowledge. If you think that's a stretch, consider how often modern advocates of educational rigour insist, like YoC, that the future is about "the three Rs and a C". Lottie Dexter is not William Tyndale, but she is in a long line of those seeking to retool the people in the name of progress.

What these initiatives also share is the fear that unless you acquire this "skill" you will be socially and economically disadvantaged in the future. It's not quite as bad as facing the rapture unbaptised, but the same millennial anxiety is there. The promoters are often not "coders" themselves (any more than your average "tech entrepreneur" is), but then the loudest advocates for the three Rs are not always numerate or literate beyond the functional minimum. But we shouldn't assume that this is just a lobby for whatever suits the needs of business, not least because "coding" does not objectively exist in the way that maths or written English does: there are many programming languages, their useful lifespan is a few decades at best, the difference between assembler and HTML is vaster than any two human languages etc. Even the argument that it's about "understanding principles" is weak as most of those principles reduce to algebra, syntax and symbolic logic. With the exception of set theory, most of this would have been familiar to schoolkids a hundred years ago.

The instrumental argument, that the future wealth of the nation depends on everyone being able to code, is demonstrably false. Contrary to popular belief, software development does not create lots of jobs. If it did, it would have done so by now. There were 1.1 million people employed in IT roles across all sectors in the UK in 2012, of which 298,000 were programmers (using the most generous definition). In other words, coding accounts for 1% of the workforce. You'll not lose money betting that this share will increase, but it will be some time before there are more programmers than public sector teachers (442,000 in 2012) or estate agents (562,000 in 2013). In fact, much of the forecast demand for IT skills is actually for intermediary roles (i.e. species of general management), rather than "coders", particularly in hype areas such as cyber-security, risk management, "Green IT", big data and cloud computing.

Unlike physical engineering, software engineering benefits from massive economies of scale and a tendency towards monopoly. We need hundreds of thousands of engineered structures across the country, but we don't need thousands of search engines or email clients, while the ubiquity of "apps" is a reflection of their ease and cheapness of supply, not insatiable demand. There will not be a Silicon Roundabout in most towns. It is also worth bearing in mind that software production is subject to the same technological and competitive forces as any production process, so there is a tendency towards deskilling as languages become more "high level" (i.e. closer to simplified English), programs write programs (i.e. AI substitutes for humans), and free or cheap GUI tools lay waste the cottage industry of Website builders.

Entry-level coding (such as Web page markup and scripting, which is what initiatives like YoC mainly push) is poorly-paid, largely because it is easily offshored. This is not "high tech" in any meaningful sense. The "well-paid jobs" in the UK that can legitimately be categorised as programming are largely species of business services, such as inhouse custom development or third-party systems integration, where the market price ultimately depends on proprietary or domain knowledge that can't be taught in schools. Programming a Raspberry Pi may be fun, but it remains less vocationally useful than knowing how to use Microsoft Excel, while the generic technology skills needed for the future - i.e. the ones more appropriately compared to the three Rs - are routinely acquired by kids outside of school via their use of the Internet and social media.

It would be easy to dismiss YoC as a stalking horse for neoliberal sweetheart deals, where privileged providers like Codecademy get to sell over-priced IT course material to the Departments of Education or BIS, but I think there is something else at work here. This is the notion of digital citizenship, on which left-of-centre critics like John Naughton find common ground with George Osborne, who says in the YoC launch video: "I want to make sure that kids in our schools are not just consumers of technology ... they understand coding ... they understand more about the world around them, and that is what education is all about".

Underlying this is the idea that citizenship is contingent, dependent on a specific ability or qualification, hence the lazy claim that "coding is the new literacy". No one is suggesting that you should be denied the vote if you are "incomputerate", any more than if you can't read or write, but there is the implication that you will be unable to "fully participate" if you lack the requisite skills. You will definitely be in the "out group", even one of the "digitally disenfranchised" (even if that phrase is really just a euphemism for "too poor to have Internet access"). I remember the 1970s BBC's series, On The Move, which tried to destigmatise adult illiteracy through the encouraging progress of Bob Hoskins' removal man. That was an exemplary social-democrat tract about what you could gain: your salvation, in other words. The "code or die" trope focuses on what you are at risk of losing, and is another example of the constant striving and performative self-improvement of neoliberal ideology: the low hum of anxiety about "outdated skills" and becoming "unmarketable".

I think there is another layer to this, which is that the tech industry is keen to encourage even more content creation by users. Learning how to create a Web page or edit a stylesheet may not be quite the liberation we imagine, but it does normalise the idea that we are responsible for continuing to "build the Web" at a time when the growth in new users is slowing. The YoC site hints at this: "We use code to build websites and apps, design clothes, publish books, make games and music, and to get the most from technology. Getting to know code is really important. It means you can be creative with computers, start your own business or boost your earning potential. It is really simple to learn and anyone can do it - not just rocket scientists". Code will set you free, but you'll have to work at it.

Whereas the digital economy pre-Web was largely about businesses creating and consuming their own content, the modern digital economy is based on the idea that everyone should create their own and consume each others. The dominant tech companies are now essentially intermediaries. Volume and monopoly allow them to charge small rents on transactions, typically via advertising and premium offerings. Though they are lionised as creators of new wonders (hence the symbolic importance of driverless cars and delivery drones), the reality is that most monetisable software is a mish-mash of acquisitions and open-source (often well-hidden), and heavily-dependent on unpaid user testing and support. Tech entrepreneurs remain first and foremost financial engineers.

No comments:

Post a Comment