In the early 1980s, I used to cycle all over London. It was my main means of travelling to work and at weekends I'd often choose to visit some far-flung corner of our magnificent capital city.
Then circumstances changed. I stopped cycling to work. My bicycle was locked up in a shed and began to rust. I used it very rarely and when on one such occasion it was nicked I didn't replace it.
Then, 15 years on, I took up cycling to work again. Paying hundreds of pounds for the privilege of joining the scrum for a bus service of questionable quality lost its appeal. For less than the price of an annual season ticket, I bought a bicycle which has lasted three years so far. At first glance, a bicycle now looks no different from one in the early 1980s. Compare this to, say, cars or computers and it's a surprise. You'd expect bicycles to move on.
It's only when you ride them regularly that you realise how much they have in fact changed. The changes are small but they make all the difference. The mechanism for changing gears, for example, works every time. In the old days, it didn't. The brakes are much more reliable and efficient. Bicycle lights are much brighter. Above all, the tyres are tougher and better designed so that punctures, the bane of any cyclist's life, have become pretty rare (even though the amount of broken glass on the street has remained constant).
The result of all these small changes, combined, is massively enhanced performance. Even through this wet winter, cycling has been a pleasure.
The moral, I think, is one about continuous improvement. While big, visible change is sometimes necessary, often performance enhancement will come from careful attention to the details. When I read Stigler and Hiebert's book, The Teaching Gap, last year, I was struck by the similarity between their description of the Japanese concept of "lesson study" and the improvements over two decades in bicycle performace.
Lesson study involves a small team of teachers planning a single lesson together. They try to make it the best lesson it can possibly be. One of them then delivers it while the others observe. Then they review the lesson, examine the students' work that emerged from it, refine it, improve it and deliver it again. Both the process and the outcome help contribute to continuous improvement.
This kind of thinking is beginning to inform the national literacy and numeracy strategies. At their core, the strategies are high volume, high-quality programmes of professional development. The content of the professional development programmes designed nationally involves a similar review process nationally to that implied in "lesson study".
This year's programmes are based on an analysis of teaching, learning and pupil outcomes last year. The five-day maths course, which has proved so popular and successful for more than 40,000 teachers, has been constantly refined in this way.
This process is having an impact at school and local levels. Growing numbers of leading maths teachers and expert literacy teachers are modelling good practice and discussing it with colleagues. Groups of teachers in schools are examining how they can refine and improve daily mathematics lessons and literacy hours. As they do so, the Government needs to learn from them and build the successful refinements into the national programmes.
Everyone in primary schools now knows what daily mathematics lessons and literacy hours look like, just as everyone recognises a bicycle when they see one.
Continuous future improvement, to build on the dramatic progress of the past two years, will come from attention to the less obviously visible pedagogical details. The changes that will result may be small but, combined, they will make all the difference.
Michael Barber is head of the Government's standards and effectiveness unit