Archives of the Future

In 2003, Roy Rosenzweig framed digital preservation around a paradox of scarcity and abundance. Two decades later, the same problem operates at a vastly greater scale and under more concentrated infrastructural conditions. Global data volume was projected to reach approximately 181 zettabytes in 2025, enough, if stored on 25 GB Blu-ray discs, to produce a stack extending some twenty-one times the distance from the Earth to the Moon. The vulnerability of digital records that concerned Rosenzweig has only become more consequential at this scale. Today's big data are distributed across servers, data centres and networks even as ownership and computational capacity have become increasingly concentrated since the Web 1.0 era. Replication, migration, modification and deletion belong to the ordinary operation of these systems. The abundance of the present may leave future historians with little more than scattered remnants of an information environment whose original scale was astronomical.

Existing preservation efforts reveal the magnitude of the disparity. Between 1996 and 2020, the Internet Archive's Wayback Machine accumulated some 733 billion web objects occupying approximately 70 petabytes, an extraordinary scholarly resource that nevertheless amounted to about 0.00012 per cent of the 59 zettabytes of data produced globally in 2020. The quantities are not directly equivalent, but their difference in scale is instructive. More important, much of the contemporary record does not exist as a bounded object awaiting preservation. Social-media records, sensor outputs, user-behaviour data, audiovisual streams and machine-generated materials are sharded, replicated, cached, recombined, migrated and purged as part of the systems in which they operate. There may be no singular authoritative copy to deposit in an archive, while the interfaces, metadata, software and computational contexts required to interpret a surviving copy may disappear before the copy itself.

Archives of the Future studies this period before the archive. Preservation is one part of a larger problem encompassing the production, transmission, migration, authentication, ownership, accessibility and disappearance of digital evidence. A record documented as destroyed, one that was never captured, one known to exist but presently inaccessible, and one whose fate cannot be established represent distinct evidentiary conditions; only the first establishes loss. BDSL documents these processes while the systems that produce them can still be observed, using field research, interviews, technical and legal records, and direct investigation of contemporary information infrastructure. The aim is not to preserve the digital present at anything approaching its original scale. It is to leave a record of how that information was created, circulated, transformed and lost, so that the fragments that remain can be understood in relation to the systems that produced them.