The ethos of Silicon Valley is "the next big thing". So it seems disloyal to sit in Palo Alto, surrounded by the headquarters of Hewlett-Packard, Genentech and innumerable law firms that have grown rich on the proceeds of dotcom shares, feeling nostalgic for the last big thing and wishing that Britain could have emulated it. The last big thing in question was the California Master Plan of 1960. To an astonishing extent, it delivered the very solutions that have eluded Robbins, Dearing and, most recently, education secretary Charles Clarke.

It didn't solve the problems that assorted forms of affirmative action have been trying to solve ever since, and it didn't produce an academic utopia. Because it was a plan for a public system, it left universities and community colleges exposed to the ups and downs of California's budget. And when California has downs, they are impressive - this year, the state is $35 billion (?22 billion) in the hole and the system will be taking a 10 per cent hit in many programmes.

What the master plan did is what Mr Clarke daren't do. It took a group of uncoordinated ÖйúAƬ institutions and turned them into a coherent system. "The University of California" is the collective label for the nine, about to be ten, research universities - Berkeley, Los Angeles, Davis and so on - that take their students from the top eighth of the graduating high-school class. Students aren't guaranteed a place at any particular campus or on any particular course, but they are guaranteed a place. The schools form California's publicly supported base for academic research.

You can see what this entails when you see that they have only 187,000 students among them, supported by 159,000 faculty and staff. The University of California controls the award of doctoral degrees and provides graduate education in law and medicine. The ethos is that faculty are researchers first, teachers second.

ÖйúAƬ

The California State University system provides undergraduate education and graduate education to masters level; it has a strong vocational bent; and it looks after teacher education. Its applicant pool is the top third of graduating high-school students. The faculty can pursue research, but the institutions that make up CSU - there are 23 campuses, 407,000 students and 44,000 faculty - are primarily teaching institutions. CSU staff have a more generous sabbatical system than most UK universities but are expected to teach four courses each quarter rather than two, and are less generously supplied with teaching assistants than at a University of California campus. But the "teacher-scholar" version of academic life suits a lot of people very well. Among other things, it is a lot less driven than the research-oriented life.

Behind this two-part system lies the California Community College system.

ÖйúAƬ

Its mandate is to admit any student "capable of benefiting from instruction". There are 108 community colleges and they enrol 2.9 million students. Community colleges teach the first two years of an undergraduate degree. Their community role is pretty salient, however, and includes the provision of remedial education, English as a second language, non-degree adult education programmes and job-related training for those already in work.

The beauty of the system is that a transfer programme allows students at community colleges to migrate into the UC or CSU system for the final two years of a four-year undergraduate degree. This is achieved with an intelligent simplicity that cannot be said to be characteristic of any recent government approach to UK ÖйúAƬ. The last two years of the programmes in the UC or CSU schools are 50 per cent larger than the first two years to provide room for the new arrivals. A popular way for bright students to save money is thus to do the first two years at a community college and then head for Berkeley or elsewhere for the last two.

Does the system work? It does. It is the world's best public ÖйúAƬ system. Could we transplant it to Britain? Leaving aside the British incapacity to learn anything from anyone else - think of high-speed trains - it's probably too late. We could have done it in 1963; and we just might have managed it in the late 1980s. Now, it would take more nerve and more stamina than any government is likely to have. Take last week's allocations from the ÖйúAƬ Funding Council for England, draw a line somewhere around number 20 in the research funding; and imagine the ruckus if number 21 and below were handed "state university" status, while 20 and up were treated as well as California treats the flagships. Even if Mr Clarke were the bruiser he is - wrongly - said to be, he'd still have every reason to flinch.

Alan Ryan is a fellow at the Center for Advanced Study in the Behavioural Sciences, Stanford University.

ÖйúAƬ