MoMA
May 20, 2015  |  Media Conservation
Preserving the Technical History of Media Works

My colleagues in media conservation have spent the last few weeks providing insight into our work at MoMA. This post will give you an idea of one small part of media conservation that is aimed at improving documentation policies related to the process history of time-based media. My role in media conservation over the past eight months at MoMA is, officially, the National Digital Stewardship Resident. This position is part of an Institute of Museum and Library Services (IMLS) grant-funded program that places recent graduates from library and archive programs in prestigious cultural heritage institutions across New York City to help those institutions find solutions for digital preservation problems.

The particular problem that the media conservation department is dealing with was touched on in Peter Oleksik’s post “Digitizing MoMA’s Video Collection.” He describes how artworks in the collection exist on a variety of media carriers, necessitating research and technical knowledge to migrate the most faithful source to a digital carrier, which allows for the extension of the work’s lifespan. He named some of the devices that are regularly used in these migrations, such as time-based correctors and playback decks. This information is hugely important to conservators because it allows them to understand why different iterations of an artwork look and act the way they do, and thus assist in maintaining authenticity in display and preservation. The information about how these migrations happened needs to be recorded in a very detailed way that will allow future conservators to understand the exact process of creating new copies of the work, down to the model name and serial number of each device used in the process.

MoMA Assistant Media Conservator Peter Oleksik working with Nancy Holt and Richard Serra’s Boomerang (1974). Photo: Peggy Greisinger

MoMA Assistant Media Conservator Peter Oleksik working with Nancy Holt and Richard Serra’s Boomerang (1974). Photo: Peggy Greisinger

To illustrate what this documentation might look like, we can turn to an example used in previous posts: Nancy Holt and Richard Serra’s Boomerang (1974). The master copy of this artwork is a two-inch Quadraplex tape, and over the past few decades it has also been migrated to three-quarter inch U-matic, Betacam SP, and a digital video file. But how exactly did this process happen? What tools did the people doing those migrations use? To migrate from Betacam SP to a digital video file, we know that a Sony DVW-A500 was used as a playback device and BlackMagic Decklink Studio was used to capture the digital video file. In most cases we either have a record of this chain, or can make educated inferences about the tools used to migrate to different formats—information that we call the “process history” of an artwork.

In our current workflow, information about process history is recorded, but not in a format that is useful for searching or doing analysis of data. The process history is generally recorded in the Conservation department’s artwork folders as unstructured text, without any concerted efforts at controlling what terms are used or structuring the data in a particular way. This means that doing automated searching of process history documentation is not possible, and also that information may end up duplicated in multiple places. Furthermore, the information is currently being recorded in programs such as Microsoft Word and Adobe Acrobat; while these programs are sufficient for some purposes, the fact that they generate files in proprietary formats is concerning for long-term preservation planning. Twenty years from now, Microsoft may not offer any support for a version of Word from 2015, and if that is the only format we have our documentation stored in, we may have a great deal of difficulty accessing our data.

Example of a process history record, written in XML. The tags, in pink, explain what each piece of data is describing. The “revtmd” prefix in each tag refers to the metadata standard, the National Archives’ reVTMD, being used to organize this record. Photo: Peter Oleksik

Example of a process history record, written in XML. The tags, in pink, explain what each piece of data is describing. The “revtmd” prefix in each tag refers to the metadata standard, the National Archives’ reVTMD, being used to organize this record. Photo: Ben Fino-Radin

Since we now have Binder, a digital repository that facilitates recording the relationships between different iterations of artworks, it makes sense to utilize this system to create a better way of recording and accessing process history records in a standardized, open-source format. We have recently finished the process of designing how this will work in Binder. To return to our Boomerang example, not only would we know that the Sony DVW-A500 was used in the creation of the digital video file copy of that artwork, but we would also know information like the serial number of the specific DVW-A500 used, and any settings that were changed on this device that might affect how the digital file looks and functions. This information will be recorded for every device used in the migration, and those records will be searchable through Binder. We have also designed the specifications for mapping the process history records to an XML (Extensible Markup Language) file. XML is a non-proprietary, text-based file format that allows information to be encoded with descriptive tags that indicate what each piece of text represents. For example, the serial number of the DVW-A500 would be encoded in a tag named “serial number,” making these records much easier for humans to read and understand.

Recording information with such specificity may seem over the top, but for media conservation it is an absolute necessity as it is capturing the technical history of the work. Imagine that, 20 years from now, Boomerang goes on display for the first time in two decades. There are two digital files associated with the artwork, and each looks slightly different from the other, indicating that one of these files may not be an authentic representation of the artwork. The conservation documentation was saved as a Microsoft Word file, but the format it was saved in cannot be opened with the 2035 version of Microsoft Word. However, because the documentation was also stored as a process history record attached to the artwork in Binder, the conservator can quickly understand the technical history of the artwork and make informed decisions about how to preserve or display the work. Properly recording and storing the process history of time-based media artworks is an important part of their long-term preservation.