It is June 2015, and the set-piece examinations that used to make this time of year a misery for students and a logistical nightmare for schools are a distant memory. So too are GCSEs, A-levels and league tables.

For the young people who started school back in 2002 the boundaries between learning and assessment are blurred. For most of their schooling they have been assessed by their own teachers, and their final matriculation exams do not feel that different from the class tests they are used to.

Downloading tasks from the national examination board’s server, they submit responses as often as they like until they feel ready to have their work marked. Results for routine, electronically-marked tasks arrive online within seconds, though it takes a bit longer for students to find out how they have performed on more complex tasks. These are still assessed mainly by human examiners, who are as likely to be based in Kabul as in Kettering .

As yet unimaginable advances in technology - and policy decisions - could, of course, make this scenario laughably obsolete long before today’s reception class pupils are 18. But it is probably safe to say that computers will play an increasingly important role in testing.

The shape of things to come is becoming visible in some of the work that test-publishing companies are doing today in 2002. Granada Learning and nferNelson, for example, have developed an online-assessment system called TestWise that schools can use to create and administer tests, score them electronically and analyse the results.

In the United States, several states are now using online testing for either diagnostic purposes or higher-stakes, end-of-course assessments. Most of these are multiple-choice tests that can easily be marked by computer. A few states, including Pennsylvania, Indiana and Oregon, have also been piloting essay-grading software. But this is dogged by problems that will not be solved soon, according to Hugh Burkhardt, director of the Mathematics Assessment Resources Service, an international project based in the Universities of Berkeley, Michigan State, Nottingham and Durham.

Professor Burkhardt points out that essay-marking systems such as e-rater, developed by the US test publisher ETS, work by measuring sentence length and the occurrence of longer words, especially those related to the essay topic.

“But what the student wrote can be complete nonsense because there is no analysis of the meaning of the writing. So the amount of marking that can be done by computers will increase steadily but it is doubtful if you will be able to get computer marking of rich human responses such as projects and so on,” says Professor Burkhardt, who has contributed to a book called Whither Assessment? to be published by the Qualifications and Curriculum Authority in February.

Over the next few years, essay-marking software will probably have less impact on the British assessment system than scanning technology. This has the potential for making the system less labour intensive and accident prone. “At present we shunt bits of paper around the country to the school, to the examiners and back to the examining boards,” says Paul Sokoloff, qualifications director at the Edexcel examining board. “But if you scan effectively, you can capture images (from scripts) electronically and move them around the country quickly.”

Right now that is easier said than done. Last year Edexcel ran a pilot project to scan and mark GCSE scripts on screen, and e-mail responses back to schools within two or three days. The experiment failed because the schools were not able to download the scanned images quickly enough. Paul Sokoloff is confident that these technical glitches will be ironed out once broadband technology becomes available nationally.

But creating an effective online-testing system involves more than giving schools fast internet access. As the OCR examination board’s experiments have shown, examiners need to adapt to the new environment. For example, it can be harder to look through a whole essay when it appears on screen than on paper. Examiners also tend be more severe when they are marking on screen.

“We don’t know why, but that seems to have been discovered elsewhere too, so that’s an issue we need to address,” says Ron McLone, chief executive of OCR. Its sister company, Cambridge International Examinations, has experimented with online testing for the Singapore government.

Once the practical problems associated with computer-based testing have been tackled, educators will have to confront the really difficult questions - such as what to test with all this gadgetry.

“Is it worth the investment that you’d have to put in, just to provide a multiple-choice test on screen?” Paul Sokoloff asks. “The answer is No. You have to use technology to make a quantum leap forward... and that’s about analysing the needs of the economy and society.”