The Art & Science of Web Design

Chapter One: Foundations

Charles Goldfarb liked to get people lost.

It was 1966. Two years out of Harvard, the young lawyer was already bored with the frustrating redundancy of preparing briefs for the firm that employed him. To burn off some energy, Charles would spend countless hours working on his hobby: organizing Boston-area sports car rallies.

As "rallymaster," he would plot courses for the roadsters on maps, then convert the courses to a detailed set of instructions. It was a game for Charles, and he enjoyed encoding logic puzzles into his crib sheets. Instead of a simple list of instructions, he would add commands like "Repeat the last six steps replacing 'right turn' with 'left turn'."

Eventually, a friend told Charles his routes were just like computer programs. "Really?" he replied. "What's a computer program?" Soon, he found that IBM would pay him a comfortable salary to write his logic-based instructions for computers, rather than driving enthusiasts. Suddenly, if you'll excuse the pun, his career took a permanent turn.

By 1969, the excitement had worn off the thrill of punch-card coding mainframes. Charles was beginning to consider heading back to the courtroom, but before he did, IBM offered him an interesting project: figure out how to apply current computer technology to the practice of law. The idea was to store legal briefs as electronic text in a database, then let lawyers query that information and recombine the results into new documents. The problem reminded Charles of the frustration he had felt years ago, sending dictated briefs over and over again to a secretary for revision and retyping-an exceptionally inefficient process.

The rudimentary text storage systems of the time were capable of storing documents and spitting them back out again-while retaining the basic formatting encoded within. But Charles found that storing the text in a database (even if that database used cardboard media) was the easy part-getting at the text and doing something interesting with it was the hard part. At first, he considered stripping all the text clean of any formatting at all, then retrieving it using simple text searching algorithms. But what if you wanted to do more compelling things than just find an occurrence of a few words? What if you wanted to get just a list of document subheads, or find all the documents written by a particular lawyer, or on a particular legal precedent? Charles faced a dilemma. How could he store the text in a database so that it was both formatted for proper output, but also could be queried in powerful ways? A search for a solution was, in fact, a lesson in publishing history.