Today, the word 'Glamour' conjures up mental images of stick-thin model girls, glossy magazines and clothes that only the absurdly rich could wear, and then only on the most glitzy of occasions.

It wasn't always thus. In the Middle Ages, learning was highly admired. Only the lucky few were able to go to the ancient universities of Europe – Bologna, the Sorbonne, Oxford, Heidelberg… – to study the 'Trivium' of grammar, logic, and rhetoric. These were considered to be the three pillars of learning, to be supplemented if you had the time and skill by other courtly arts such as music and fencing (with a sword). All three pillars were about the formalisation of discourse: grammar for its linguistic rules; logic for its laws of inference; and rhetoric for its heuristics of effective persuasion.

Grammar covered both syntactic and lexical analysis, and involved memorising huge chunks of Virgil and Homer. Logic meant reading the philosophers Aristotle and Plato, and Rhetoric meant the speechwriters such as Cicero. In fact, whatever you wanted to learn, the prescription was to read the classics. One would like to hope that today we have more than one approach to offer our students and clients.

Rumour of this arcane wisdom spread dimly into the alehouses and taverns of the hinterlands of the university cities, the very word Grammar sliding into a hazy 'Glamour' of splendour surrounding the rich, powerful, and university-educated. Indeed there were magical connotations, as in Gramarye, a book (and possibly also a faraway land) of sorcery. It's our job not to appear as a secretive and erudite priesthood poring over books of incomprehensible specifications (and it's perhaps worrying that hierarchy means 'rule by priests').

Learning does not, perhaps, have quite the same allure in a world of instant communications and competing attractions. But the old magic is not quite dead. It would indeed be nice to find a way of specifying systems precisely in a grammar of totally exact meaning, with a logic that could not be escaped, and in a presentation format – a rhetoric – that really persuaded. Maybe the Trivium is not as remote as it seems.

The search for the perfect language is – didn't you know? – centuries-old among philosophers, for then you could translate any human tongue exactly into any other, via this wonderfully unambiguous but unattainable intermediary. Its 20^th century version was the attempt to devise a uniform representation for any natural language text, for use in machine translation: a novel application of an ancient and utterly misguided idea. Italian linguists have a proverb: translation is treason – for every attempt must betray some part of the original meaning.

A parallel quest was for a complete Encyclopaedia, a codification of all knowledge in all domains: a modest intellectual goal, you'll agree. Diderot in 18^th Century France set out to describe everything from the most modern agricultural machines to the different styles of music; but even as the first of many planned volumes was published, he realized that it was out of date. The third Ming emperor of China was more determined. Between 1404 and 1408, he set over 2000 scholars to work to produce the Yung Lo Ta Tien, 11,095 handwritten volumes "executed on fine white paper ruled in vermilion and bound in hard boards covered with yellow silk", according to Robert Twigger, The Extinction Club, Penguin, 2001. It was too large to be printed, so no more than one or two copies were ever made, and most of the volumes have been lost or burnt, though some 370 remain as curiosities in the world's museums and libraries, 15^th Century shelfware.

We can quite understand why researchers today dream of creating a formal method (supported by tools) that ruthlessly enforces complete consistency. But alas! technological change defeated Diderot's Encyclopédie, and human vagueness defeats all attempts at putting precision into people. Ben Kovitz [RE'02 Industrial Talks, 103-110] is right: formality is the endpoint of our work, the ultimate goal, not something we can apply directly to the ambiguous place where people live and work. Requirements Engineering is and always has been about starting from where our stakeholders actually are. Al Davis and Ann Hickey have some stern words on this subject for researchers [Requirements Researchers: Do We Practice What We Preach?, REJ (2002) 7:107-111].

If we forget this awkward premise, our discipline will become as 'Trivial' as the whole of mediaeval scholasticism came to be seen; and our codifications of knowledge will be about as relevant to working life as the only extant bit of the Encyclopédie, Volume 1 of 10,000, Aa - Aardvark. Remaining 9,999 Volumes Coming Soon! Or whatever.

© Ian Alexander 2002

Home Page