Three Test cases have been used so far to evaluate existing preservation strategies as well as test our own proposed strategy. These test cases are,

1. Verbum Interactive - an interactive digital art magazine on CD-ROM;

2. Eurovision - a video artwork on VHS; and

3. The Elements - a new media artwork by Nam June Paik comprising physical objects combined with video monitors and audio speakers in a particular layout.

Verbum Interactive

About Verbum

Verbum Interactive is described as a digital art and design CD ROM Magazine. In fact it is claimed to be the world's first fully integrated multimedia magazine. Verbum Interactive combines text, sound, graphics, animation, talking agents, video and music, accessible via a point and click interface. Two CDs make up the magazine: the first CD features stories with on-screen and printable text, sound, video clips and sample multimedia files. The second CD is an interactive panel discussion between six industry leaders on the state of the multimedia industry.

The Problem

Verbum Interactive was published in 1991. This edition was developed on a Macintosh Classic and required very specific system configuration to function. When the Macintosh Classic became obsolete so did this edition of Verbum Interactive. The contents of the two CDs became inaccessible and this historically important inaugural interactive multimedia magazine became lost despite the fact that the physical storage medium, the CDs, have not deteriorated.

As no stepwise migration had taken place before the contents of the CD became inaccessible, migration could not be performed. Reinterpretation was also not an option since there was no information regarding the structure of the work or links to the various components of the work to enable reconstruction. This highlights the role that metadata can play in preservation. If preservation information had been recorded, including the location of the various components of text, images and other multimedia files and more importantly the information regarding the structure of the work, then it may have been possible to re-create the work on the latest platform with the original content. However the lack of metadata meant that the only strategy available to us was to attempt to play the CDs on an emulated Macintosh.

The Preservation Strategy

The emulator which we used was the Basilisk II, an Open Source 68k Macintosh emulator which enables you to run 68k MacOS software on any computer. However, a copy of MacOS and a Macintosh ROM image was required in order to use Basilisk II. We were able to access approx. two thirds of the original content of Verbum using Basilisk II. Most of the problems that arose were in relation to the first CD. The second CD appeared to be fully accessible. The problems we encountered included:

• When the emulator was set to emulate a MacOS with optimal setting for the two CD then the animated graphics interfered with the emulation software and the emulator froze. With less than optimal setting many of the graphics in the CD would not work.

• Certain parts of the first CD could not be accessed because the emulator is not a perfect or complete emulation of the MacOS system.

An interesting point to note is that Verbum Interactive does contain some information that can be used to reconstruct the CDs, but because the data is stored on the CDs the information is also rendered inaccessible because the CD is inaccessible. Thus it makes sense to store the preservation metadata separate to the actual content and to link or point from the metadata to the content.

Eurovision

About Eurovision

Eurovision is a video artwork by Linda Wallace, completed in March 2001, recorded as a PAL format video of length 19.2 minutes. Using video composition software, Linda Wallace has combined four sequences from the Eurovision song contest with footage from Bergman and Godard films, a Russian documentary about the space race from the 50s/60s, images from the Louvre and icons of the Madonna, with a background music track and overlaid text - in a multi-framed video stream.

The problem

Currently there is no problem accessing this work via any VCR which supports PAL. Ancillary contextual information (metadata) is also available via the associated web site. However with video cassettes fast becoming superseded by DVD players and the indifferent durability of magnetic tapes in the long run, there is a good chance that Eurovision will not be viewable in years to come unless refreshing (transfer to a better physical storage) and migration (reformatting to a high quality, standardized format e.g., Motion-JPEG or MPEG-2 (HDTV)) steps are taken.

The Preservation Strategy

Besser and Wactlar et.al. have both proposed similar approaches to the preservation of moving image materials. Both groups propose the digitization of the film/video in a high quality standardized format (e.g., Motion JPEG or HDTV-quality MPEG-2) and storage on RAID storage systems, which are periodically refreshed. In addition, both assert that the recording of associated metadata is essential for ensuring long-term accessibility to moving image content.

MPEG-7, the Multimedia Content Description Interface, is a standard developed by the Moving Pictures Expert Group (MPEG), a working group of ISO/IEC. Its goal is to provide a rich set of standardized tools to enable both humans and machines to generate and understand audiovisual descriptions which can be used to enable fast efficient retrieval from digital archives (pull applications) as well as filtering of streamed audiovisual broadcasts on the Internet (push applications).

Although MPEG-7 provides a very comprehensive set of description tools, to enable the creation of (XML) descriptions of multimedia content, it does not support the preservation needs of moving image content, which requires metadata such as:

• Contextual information provided by the artist e.g., the meaning behind the work, how it was created, how they would like it displayed etc.

• Types of reformatting, refreshing, preservation actions permitted by the artist/owner;

•  Details of refreshing and migration events.

Some of this metadata may be provided in the future by MPEG-21, the Multimedia Delivery Framework which aims to provide a framework to enable transparent and augmented use of multimedia resources across a wide range of networks and devices to meet the needs of all users. However until the MPEG-21 standard is final and complete, the extent to which it will satisfy the preservation metadata requirements of moving image content, will remain fuzzy. The Library of Congress Audiovisual Prototyping Project has also developed a draft set of extension schemas to use within METS for encoding preservation metadata specific to audiovisual content. The additional metadata is primarily technical metadata (specific to audio, images, video and text e.g., sampling rate, compression codec etc.) for both the current object and its source plus details of reformatting and migration events.

Until MPEG-21 is available, the optimum approach to capturing the preservation metadata for Eurovision would be to use the LoC's Audiovisual Prototyping Project schemas for video but further extended with some additional metadata specific to artworks, such as the artist's intention and perspective, technique, tools used and their views on how the artwork should be installed and screened and which preservation strategies are permissible. The Variable Media Questionnaire could be used to provide this additional set of artwork-specific metadata elements. The complete set of metadata elements and links to the various digital components would be stored within a METS container.

The Elements

About The Elements

The Elements by Nam June Paik is a typical example of a large class of new media artworks. In The Elements, Nam June Paik has set six television monitors into an oriental red lacquer cabinet and plays video footage featuring early Hollywood film stars alongside avant-garde artists and rapid sequences of images of the natural world.

The Problem

New media artworks such as The Elements are notoriously difficult to install or transport. They rely on a large set of fragile physical and digital components with complex interdependencies and physical, spatial and temporal relationships which need to be satisfied in order to maintain the integrity of the artwork according to the artist's wishes. The problem is one of ensuring that all of the metadata which is required to recreate the artwork has been precisely recorded and that the various components do not degrade or alter with time or during refreshing or migration steps.

The Preservation Strategy

For complex mixed-media artworks such as The Elements which are composed of multiple audio and video channels, images, text and even physical objects combined in some spatio-temporal structure, we propose an approach based on a combination of metadata and migration to high quality, platform-independent, standardized formats. In order to ensure that all of the relevant metadata is recorded, an analysis of the workflow from content creation to acquisition to exhibition to archival to reuse and preservation was carried out. Figure 1 illustrates the typical life cycle of a new media artwork and the kinds of metadata which are acquired/modified at each stage. Based on this analysis of the content and metadata workflows and the proposed preservation approach, we were able to determine an optimum metadata schema to support the preservation needs of new media artworks such as The Elements.

A summary of the key extensions to METS are listed below:

• Descriptive Metadata
  •  Dublin Core Metadata Initiative
  • Guggenheim Variable Media Questionnaire
• Administrative Metadata
  •  Technical schemas proposed for use in the Library of Congress Audiovisual Prototyping Project
• Structural Metadata
  • SMIL 2.0

Fig. 1. Typical Metadata Workflow associated with Multimedia Artworks

Once the schema had been designed, the next step was to develop actual metadata input tools to facilitate the recording of preservation metadata, at the points of creation, acquisition, reuse, migration etc. The next section describes the metadata input tools which we've developed specifically to support the preservation of new media artworks but which can be applied more broadly to complex mixed-media objects from any domain.