Standard questions

‘The nice thing about standards is that there are so many of them to choose from.’ (1)

So said Dutch computer scientist Andrew S Tanenbaum – but as a professor of computer science, he would be aware that standards have played a crucial role in the evolution of one of the biggest phenomena of the past decade: the internet.

There has been a proliferation of standards on the net – from ‘http’, the basic protocol that enables web pages to be universally accessible; to mark-up languages such as HTML, XML and XHTML used in web pages; to wireless internet protocols such as 3G and GPRS. And there are enough organisations developing and maintaining these standards to make the world of boxing drool – including the Internet Engineering Task Force (IETF) (2) and the Internet Architecture Board (IAB) (3), which have as their organisational home the Internet Society (ISOC) (4), and the World Wide Web Consortium (W3C) (5).

In one sense, the internet is only about standards. For example, the only reason you are able to send an email to a friend is because a standard has been adopted that specifies the format of your message, how you should indicate who you want to send the message to, how the message should be packaged up and delivered, what to do if the other person’s mailbox is full, and so on. Every aspect of the route your message takes is completely specified. It works because the software you are using adheres to the standard; the email servers that carry your message adhere to the standard; and your friend’s email software adheres to the standard.

The existence of this email standard – and, consequently, the ability to communicate a message to somebody without concerning yourself with the software they would use to read it – was sufficient to make email the fastest growing and most widely used of all the internet tools and software.

The fact that the internet was built on standards also helped in the development of the worldwide web. Its inventor, Tim Berners-Lee (6), developed his software to show pages of information to users in such a way that they could move from one document to another, using links. However, he was able to focus on the specifics of his own software because he could take for granted the underlying protocols of the internet. These protocols allowed computers to communicate with each other over a network, so when Berners-Lee built his software to serve up his new ‘web pages’, he didn’t need to develop software to handle the underlying communication of the data.

But the benefits of standardisation are not confined to the internet. Take railway tracks – prior to 1845 there was no agreed standard in Britain on how far apart tracks should be placed. A government-appointed commission deliberated over whether the wider spacing favoured by Isambard Kingdom Brunel (7) was preferable to George Stephenson’s narrow one (8). During tests, one of the narrow gauge trains fell off the track and Brunel’s engine was more economical than a second narrow gauge locomotive that did stay the course – but preference was given to the narrower track.

The adoption of this standard got rid of the inefficiency of having to transfer goods and passengers from one train to another halfway through a journey because the tracks didn’t meet. As with our modern-day email, the issue was less about which standard was adopted and more about the adoption of any standard, making much longer unbroken journeys possible.

So standards have the capacity to cheapen production costs – and as modern manufacturing emerged and became mass production, standards became more important. Mass producing products on an assembly line requires components that are consistent and always of the same dimensions – otherwise it becomes difficult to devise machinery that can make those parts. In cases like these, mass production without standardisation is simply not possible.

An example of this in the USA is the eventual adoption of a standard screw size – the Sellers screw – across manufacturing, from Singer sewing machines to the Pennsylvania railroad. ‘By the end of the century’, writes James Surowiecki in Wired magazine, ‘the Sellers standard was effectively universal in America, and in 1901 it was adopted by Europe at the International Congress for Standards and Gauges’ (9).

In his Wealth of Nations (10), Adam Smith argues that one reason why increasing the division of labour in a production process yields greater productivity, is because once a task has been broken into many smaller tasks, it is possible to add more machinery to perform some of the tasks. Similarly, standardisation brings productivity gains – the cost of producing a standard screw, for example, will be far cheaper because it is easier to mechanise than producing many different sizes, and those who use standard-size screws benefit from cheaper components.

But adopting standards is not universally regarded as a good thing. Some see standards as a curb on individuality, in the same way that a division of labour seems to turn humans into part of a machine. Even the adoption of the Sellers screw met with opposition.

‘Sellers’ proposal seemed rational’, explains James Surowiecki . ‘But many machinists viewed it as a threat to their way of life. They saw themselves as craftsmen. While they built machines designed to mass-produce goods, the machinists did not use mass-production techniques themselves. Take gun-making: in the early part of the century, gunsmiths custom-made muskets to the specs of each buyer. By the time of the Civil War, rifles were being churned out by the million. Interchangeable parts helped win the war, but in the process they reduced the American gunsmith to little more than a maintenance man.’ (11)

We hear echoes of this opposition today, when those who work in IT claim that individuality is stifled by the imposition of standards. But whether in the software houses of the twenty-first century or the mass production of the nineteenth, standards have to be a good thing, saving both time and resources. Surely there are better outlets for expressing your individuality than the taper of a screw?

However, the argument that standardisation can stifle innovation is far more interesting. Berners-Lee was truly innovative in the way he devised his hypertext browser, and a server that could deliver hypertext documents. At the same time as he developed his software, many others were working on similar applications – but nobody had made the leap of the imagination that Berners-Lee did, to see that a hypertext system needed to be able to work across networks.

But the internet bubble that his invention prompted is not so innovative. In technical terms, hardly anything fresh and original has come out since Berners-Lee’s web browser. If anything, the internet bubble was a consequence of a lack of innovation. At a time when there was little in the way of invention, everybody latched on to the only big idea in town. In more imaginative times, there would have been plenty of new ideas to develop and new techniques to explore – plenty for the next generation of entrepreneurs to get their teeth into.

In the 1990s, the internet boom took off, not on a wave of innovation and creativity, but because people had little better to do. Twenty- and thirtysomethings with few ideas of their own and little technical skill rode on the back of Berners-Lee’s invention. His innovation was forced to bear the weight of the lack of new ideas, as everybody clambered on board.

The consequence of this lack of innovation is that standards proliferate. Where in the past it took a government-backed committee, a locomotive competition and the passing of decades before Brunel’s wide gauge finally gave way to Stephenson’s narrow gauge, today’s standards are handed out by academics and committees on an almost daily basis. In a climate of confident invention, most companies would be jealously guarding their data formats so that they could lock in their clients.

Today, we have a business culture that avoids risks at all costs. Few companies nail their colours to the mast on some new technique or format without getting the support of at least a handful of other companies and possibly a committee (12). This is partly because they’re worried that nobody will follow them, and partly because there’s usually a backlash against companies that adopt their own proprietary standards, since standards have today become an end in themselves.

So what are we to conclude? Standards are a good thing, saving us time and effort and raising productivity. But we can’t help feeling disappointed when the development of standards, and the act of conforming to them, is presented as innovation. When we’re content simply to produce standards, something has gone wrong with vision and innovation more broadly.

Of course standards don’t stifle innovation. But standards that emerge from competition in the market at least imply that there is some underlying innovation demanding our attention. Today, market-driven innovation has given way to sanitised implementation – with industry being drip-fed technical advances from on high in the form of new standards.

Perhaps we need a bit more of the maverick non-conformism of the old days.

Mark Birbeck is managing director of x-port.net Ltd, and was responsible for building the ePolitix.com political portal. He is co-author of Professional XML and Professional XML Metadata, both published by Wrox Press.

Mark Birbeck is introducing the spiked-seminar ‘Who sets the standards for the next-generation internet?’ on 26 February 2002. For further details of spiked-seminars, email Sandy.Starr@spiked-online.com

(1) Andrew S Tanenbaum, Computer Networks. Buy this book from Amazon (UK) or Amazon (USA)

(2) See the Internet Engineering Task Force website

(3) See the Internet Architecture Board website

(4) See the Internet Society website

(5) See the World Wide Web Consortium website

(6) See the biography of Tim Berners-Lee on the World Wide Web Consortium website

(7) See the biography of Isambard Kingdom Brunel on the Brunel University website

(8) See the biography of George Stephenson on the Spartacus Educational website

(9) Turn of the century, James Surowiecki, Wired, January 2002

(10) Adam Smith, Wealth of Nations. Buy this book from Amazon (UK) or Amazon (USA)

(11) Turn of the century, James Surowiecki, Wired, January 2002

(12) See, for example, Sun, BEA and Microsoft work on Web Services standards, it-analysis.com, 7 February 2002