Prompt's TechBlog

Yesterday I tried to get Brian Eno's Generative Music running on my PC. It's a program that replays Eno's music and regenerates it within the parameters he's defined so that each performance is unique. The album was issued on a floppy in 1996 and was compatible with Windows 3.1 and Windows 95. It doesn't run under XP, though, which means that the music sequences Eno programmed are effectively lost to me and probably most others who don't want to tinker with the program.

More and more of our business and cultural data is stored digitally, and that means its lifespan is as short as the software and hardware it runs on. In 1986, the Domesday Project aimed to create a digital record of then-modern Britain. Barely a few years later, the BBC computer was obsolete and the video disc players were hard to find in fully working order. The data was effectively lost, until a massive rescue operation a few years back. Compared to the print edition of the 1086 Domesday Book, which is still available at London's Public Record Office in Kew, the digital version didn't last long.

Now the National Archives and Microsoft have teamed up to try to preserve digital documents. Microsoft has of course caused many of the problems with its proprietary and incompatible file formats, used to force people to upgrade software. And rather than work with the existing Open Document Format for Office 2007, Microsoft specified its own format and had it ratified as an open standard, so that it wouldn't have to cede too much control.

The deal will provide National Archives with an operating system that can run different versions of Microsoft Windows and Office applications from the same PC. The idea is that the staff will be able to swap between the operating system and application version they need to access any data.

Is this the right solution, though? Is this not perpetuating the lock-in by ensuring that this data will always be dependent on Microsoft software?

I wonder whether a better solution would be to convert data from obsolete formats into current open source formats. If the source code and format documentation is archived too, it will make it much easier for future generations to recreate code that can open files. The code can be adapted for different platforms without breaking any copyright or licensing restrictions. The software to access the data can be copied as widely as the data itself, which would ensure everybody could access it.

Clearly hardware changes will continue and even open source file formats can obsolesce. Perhaps the simplest thing of all, certainly for documents and other data files that don't depend on interactive features, is just to print them onto archive quality paper and keep them somewhere dark and dry.

Labels: data formats, microsoft, national archives