The metric system: a case study in technical standard setting

I have just finished reading The Measure of All Things : The Seven-Year Odyssey and Hidden Error That Transformed the World. It is a fascinating case study in how international standards are set.

As of 2004, the only countries in the world whose official system of measures is not the metric system are Myanmar (formerly Burma), Liberia and the USA. It is a fantastically successful international standard. In the US, Congress passed the Metric Conversion Act of 1975, calling for voluntary conversion. Amendments to the Act in 1988 designated the metric system as the “preferred system of weights and measures for United States trade and commerce.” In the US it is legal to use but not mandatory.

Before France’s revolution, its’ academy, very much an analogue of England’s Royal Society gained prominence. It was the age of rationalism and the scientists were pushing for measures which simplified scientific calculation. Think of them as the geeks. At the same time the King was worried about unrest due to lack of grain. Think of him as the suit. France had a babylon of local measures. Each city had measures morticed into the town hall walls. There would be one for a barrel, one for measuring grain, one for measuring cloth and so on. There were hundreds of measures for each item, totalling thousands of measures across France. It was very difficult to create contracts or conduct trade. As each market had its own measures, it made it difficult to compare prices between markets. The local aristocracy charged a fee to use the local measures. As a result each market was a monopoly.

The King was interested in a uniform standard. Nothing more. The idea of a completely arbitrary standard was anathema to the Academy, who thought that it should be based on an invariant, like 1/10000000 the length of the meridian from the North Pole to the Equator. Why this? They already knew this would give a measure very close to a length measure already used in Paris. The King reluctantly agreed and a surveying mission to measure the meridian from Dunkerque to Barcelona. The idea was to measure part of the meridian and then extrapolate the result, based on the idea of the Earth as a regular ovoid.

The two chief surveyors were Delambre and Mechain. In the seven years it took to complete the survey, revolution and counter revolution took place. At the end Napoleon was in charge. Though there were changes of government, each government wanted to advance trade. The new idea of nation engendered by the French Revolution needed means of defining the nation. Indeed the metric system has often been adopted when countries have become independent (e.g. India and Chine).

Ultimately it was found that:

1. The Earth is not a regular ovoid, so the extrapolation was impossible
2. That plus Mechain’s fudging of his data, caused the metre to not be 1/10000000 of the quarter meridian.
3. The metre turns out to be an arbitrary standard afterall.

Nevertheless the initial geekiness around the metre, and Napoleon’s successful conquest of much of Europe was enough to get the metre rolling. The rest of the metric system rolled along with it.

Though scientists quibbled over the measurement of the metre, economics was driving its adoption. With standards in place it was possible to trade between markets in a country and then markets between countries.

Once there was a critical mass the network effect took over. Countries started adopting the metric system because other countries had, and the standard would enable trade.

The scientists of Britain, USA and the colonies were not invited to the first metric convention. The hostility to the metric system caused by that fateful act ensured the metric system would not be used in Britain, the Commonwealth or the United States for almost 200 years. Finally with the creation of the European Common Market in 1970, Britain was concerned enough to act. It adopted the metric system and the remaining non metric countries such as Canada and Australia followed suit. The latter happened in my memory. I remember having inch rulers in Grade 1 and centimetre rulers in Grade 2.

The United States had standardised on imperial measures early in its history. Because it already had a standard it also had all the benefits that come from one. It is only now, in the age of globalisation, which is really a word for pervasive international trade, does the United States suffer. Trading partners insist on metric measures. Industries such as Car manufacturing, which rely on parts made all over the world, are already completely metric. Now almost 80% of Americans know of metric units. It is expected that the metric system will continue its creeping takeover of the USA.

A repeating theme in adoption of the metric system is its rejection by adults. In the Benelux countries who were the first to permanently adopt the metric system , it was found the only way to achieve adoption was to teach it in schools and wait two generations until the majority of the population understood it. The old measures were anthropomorphic. The foot was originally the King’s foot. Some human measures, like your weight and height feel more natural in the old measures. Everyone knows that a male over 6 feet is considered tall. I am 5’11”.

So, what are the lessons learned for introducing a new technical standard?

1. The standard should appear to have technical merit and appeal to geeks, even if it secretly doesn’t
2. The standard will need an economic rationale for adoption
3. The standard should appear to be neutral and not favour any one party
4. A meeting representing all should be held to agree on. Those not represented will inevitably resist it.
5. The standard needs an exact specification. The metre has been revised three times and its specification tightened.
6. The standard needs a standards body.
7. The standard needs easy availability of implementations. (The French produced millions of metre rulers each year)
8. Compliance needs to be audited and enforced.
9. Once the network effect kicks in the standard becomes viral and achieves dominance
10. The old standards, though rarer, will persist for a long time
11. A standard that works well enough is very hard to replace. E.g. the US imperial system

A great example of a well executed standards process is J2EE. A great example of a bad one is Web Services, where almost every one of the above rules was broken.

A final characteristic of successful standards is there permanence. It was discovered only a few years after the introduction of the metric system that the metre was flawed in terms of what it was supposed to be. A platinum bar was created which became the first metre. Then when the Germans complained that the ends were scratched, another one, created by International committe over years was fashioned. The current definition is 1/299,792,458 of the distance travelled by light in a vacuum in one second. Each refinement has simply been a better specification of that original flawed platinum bar. The length of that bar will likely live on for Millennia.

Doubtful? The 60 based time and angle system (60 seconds in a minute, 60 minutes in an hour, 60*6 = 360 degrees in a circle) originated in Babylon 4000 years ago. When the French tried to reform time into a decimal day they failed. I wonder how long TCP/IP will be around?