Sunday, April 8, 2007

Microsoft Office v. 2100?

Those of us who have an interest in technology and want to build widely adopted products will do well to consider the issue of longevity. Building products with a long-term view is a good practice in itself; but also thinking long-term yields design requirements that tend to be beneficial for the overall product.

As an example, consider the Domesday Book, a comprehensive survey of England, commissioned by William I of England, and completed in the year 1086. The book still survives and can be viewed in Britain's National Archives. Today, more than 900 years later, the book is readable, understandable, and provides valuable historical, social and economic insights to historians and scholars. Inspired by the Domesday Book, the BBC undertook to create a similar survey of England in 1986. The contents were stored as 12-inch video discs (remember those?) which are now obsolete. The digital version of the Domesday book lasted all of 20 years!

Think about how much of the information generated in the last 40 years or so is already lost to us. Data created in some proprietary format for which the program is no longer available, or stored in obsolete media which can no longer be read. Clearly we generate a lot more content that we did before, and perhaps not all of it needs to be preserved. However, since there is no way to distinguish what should be preserved from what should not (and who gets to decide this anyway?), they are treated identically. There are groups, such as the Digital Preservation Coalition, based in the UK (started as an attempt to save the 1986 version of the Domesday book as their top priority!) that are tackling the issue. But a lot more remains to be done.

Even more important than preservation efforts are design principles that allow for the creation of long-lived content. Products designed with longevity principles in mind are the only proactive and scalable way to solve the problem. Preservation will always be a reactive and expensive fall-back option and will never offer a complete solution to the problem.