what we remember is sometimes not the searchable content. In these instances we narrow the search space through circumstance reconstruction—a kind of semantic way-finding to the content… “something from 2004 when Mum came to visit, so it must have been August and it was a picture and it would have been….” Again, Apple’s Time Machine in Mac OS X Leopard explores this, giving you a snapshot in time of your files. This is an appealing idea.

A lot of human information interaction is serendipitous, based on vague, ill-formulated, semantic associations not clear on text and numbers, and enacted as browsing, encountering, and being reminded— not explicitly remembering. A text-search string still does not find a figurative image, and file metadata are volatile. But reconstructing context is a powerful memory-jogger bringing back the abstract textual that goes with the recognized visual.

Search will also need to return results that cut across different media. Google’s Universal Search, which provides results from video, images, new, local, and book search, is a step in this direction. Yahoo!’s OneSearch does this nicely for cell phones. Ask.com does it too, but prettier.

The world is waiting for the designer who can (re)create and implement the memory palaces and mnemonic techniques used by renaissance scholars and described by Frances Yates in The Art of Memory.

3. Data is dynamic, not static. The great promise of an archive is to assure long-term access to information. That sounds like

stasis, but it isn’t. To be effective over decades, archival systems need to migrate data from disk to disk, and in some cases, emulate the environments of the applications that use the data.

In considering personal data storage, we need to consider the easy migration of personal data from one location to another. But personal and social data are always evolving; they are not stable. Formats change, data migrates between storage methods and places, and security and access methods evolve. Smart organizations are looking to support users in their understanding of the consequences of that volatility. Services are beginning to take on the responsibility of educating users as well as funding research into data migration and fighting against format obsolescence (often by supporting current as well as legacy formats).

Digital rights management schemes that allow limited access today may fail in ways that allow no access tomorrow.

For designers these considerations may lead to uncomfortable practices. Refusing to innovate in favor of traditional practices and technologies; sticking close to the file system rather than adding a layer on top; and avoiding the unique in favor of the conventional as a way to support future users and avoid evolutionary dead ends all go against the desire to improve on past practice.

4. From personal to social data.

Archives sit at the boundary between public and private data. Data that was once private may, through an archive, gradually be made public. That presents new opportunities and challenges

the digital environment.

One opportunity is in cataloging, which is expensive for both institutions and individuals. When the individual is overwhelmed with too much content to name, tag, sort, and store, we could always harness the crowd, get the group to tag and organize. Crowdsourcing and services like Amazon’s Mechanical Turk harness human intelligence to solve problems that computers find hard—like tagging and organizing and storing. Archiving is a collaborative practice, and it is going to become ever more so.

But this solution brings up another issue we need to keep in mind: Who becomes responsible for the content created through a collaborative enterprise, and how are ownership and responsibility for that content conceived of by the service providers? An article in Wikipedia is distinct from the contributors who created it, but if a photo that has been collectively tagged in a photo-sharing site like Flickr “belongs” to an individual who subsequently leaves Flickr, what happens to the content? Many people are crushed when the comments they have made on blogs disappear because the blog “owner” stopped maintaining the blog.

Relying on social approaches to archiving may be a practical necessity, but open archives must be built to withstand and respond to a wide variety of attacks, not only from individual malware authors, but from political partisans, abusers of copyright law, and even governments that wish to control access to historical records.

The Society of American