This afternoon, I would like to present some of my personal observations on the promise of UPF from the perspective of a conservator of works of art. This talk is not going to provide a comprehensive list of the benefits of UPF, which I think are numerous and which I am sure Dave, Thom and Mary are prepared to enumerate.
By way of further introduction, I am an art conservator, specializing in the conservation of works of art on paper and photographic materials. Recently, I helped found the Electronic Media Group within the American Institution for Conservation and I’ve been following the UPF project over the course of the past few years. This talk really assumes some basic knowledge of UPF, but hopefully if you have missed the previous UPF-related talks at past AMIA conferences, this talk will not leave you lost since its not technical in the least.
For conservators, especially fine art conservators, "integrity" is a word that gets used a lot. When applied to a cultural object like a sculpture, a painting, a book or any other piece of material culture when we conservators say "integrity" we mean whether or not the object is intact both visually and contextually. In other words, not only does the object look the same but also it retains the more subtle, characteristic markers that place it in a certain time, place and culture. These markers can be evidence of original use, past display or methods of manufacture.
From my perspective as a conservator, and based on my understanding of UPF, I feel the implementation of the UPF approach has tremendous advantages when it comes to maintaining the "integrity" of digital culture.
More on the word "integrity:" The mission of the conservation profession is the preservation of cultural property for the future. This sounds very straightforward and, in practice, there are standards, norms and convention to guide us in this mission. For example when we are dealing with a preservation problem relating to a 19th century canvas painting, there is a significant body of expertise and traditional practice to draw upon.
However, conservation implies the preservation not only of informational value but also the preservation of the original artifact in a manner that does not corrupt or undermine the links between the object and the culture that created it.
In the digital world, the challenge here is obvious: We are so close to the creation of digital cultural material that it is easy to make changes in the name of preservation that can undermine the value of digital work for future generations. Except possibly for DaVinci’s Last Supper, which has received nearly continuous restoration from basically the moment of its completion, we have rarely had a preservation imperative a arise so quickly after original manufacture, especially on such has large scale, as we do with digital materials.
Relative to the other materials that tend to find their way into museums, archives or libraries, we will not have the benefit of a tradition of care and maintenance that will guide our actions when it comes to digital works. Though we need to act, the compressed cycle between the manufacture of digital artifacts and the almost immediate imperative to preserve these same artifacts should not lead preservation decisions based on expedience. A particular concern of mine, is that any preservation strategy account for the fact that a digital work is not only information, but potentially a piece of our shared material culture. Given that, we should at least consider how future generations will come to place a value a particular piece or collection. And by generations I do not mean a couple of generations measured by the computer / video industry, but for generations measured in human terms.
We can be pretty certain that future generations will value having access to an original old master painting, or an original print for a feature film. This certainty leads us to make certain preservation decisions, and we dedicate a fair amount of resources as a culture to make sure that the old master painting, or the film, are handed down.
However, we are new to the task of being custodians for digital creative work and we have difficulty assessing where the value of a particular digital creative work lies. Does the piece have strictly informational value, in which case a surrogate or copy might make an appropriate preservation vehicle? Or does this collection of D3’s lets say or MPEG 1’s, comprise a more substantial piece of digital culture that merits a more well-conceived and respectful preservation approach?
As we ponder this question, I want to further complicate the picture with some lessons gleaned from the study of art conservation over time:
First: The implementation of any preservation initiative inevitably reflects our own contemporary cultural context and biases. Evaluations of the intrinsic value and importance of cultural material change over time. Our view of the digital-based moving images we are creating today inevitably is not going to reflect how that work is assessed in the future. For example (in what may be an apocryphal story), when Alfred Stieglitz brought the first Cezanne watercolors into the United States he claimed them at customs as artwork. Upon inspection, the customs agent determined they were in fact "stained paper." And that’s what they were at the turn of the century, stained paper. Today, of course they are "Cezannes." I suspect there is a lot of digital "stained paper" out there today...
Second: Another lesson is that the good faith preservation work we do today may be considered overly intrusive or misconceived by future generations. All conservators with any treatment experience would love to take back many, many treatment decisions. A sort of funny, at least I think its funny example of this gets played out on nearly every episode of the Antiques Roadshow (appropriately, a WGBH production) where an early American piece of furniture gets evaluated by the guy in the snappy suit. He builds up the piece until you can see the dollar signs lighting up in the owner’s eyes. But then he drops the hammer, saying that if the piece had its original finish it would be worth a fortune, but now its only value is as a curiosity, or a study piece. Usually the owners replaced the finish, and after the camera cuts away you imagine them slinking off, crushed and despised.
Now, no one questions that the original finish was deteriorated to an extent that would be considered highly undesirable and even ugly by the original furniture maker, but in a contemporary context, that finish has tremendous value as an unbroken link to the past. This contemporary value we place in the disfiguring, but nonetheless original, finish speaks of the generally non-interventive approach to preservation that is a distinct contemporary trend in the professional practice of art conservation.
Last week I was talking with Jill Sterrett, Conservator at the SFMOMA, a museum that accession artists’ work stored on CD-ROM, web pages, video art, installation art. Speaking of this new media she said "Restoration is perhaps at odds with modernity." The point here is that less is more: the less we do to alter the inherent attributes of an original work, the better job we are doing as custodians.
So lets bring all that context and apply it to UPF: As a neutral container, UPF envelops the original digital creative work without altering its native state. The potential of a viable UPF is tremendously liberating since it substantiality removes the hazards that have traditionally hampered the preservation of conventional media. In other words:
1. UPF will provide for the maintenance of both the informational value and the cultural context of a of digital creative work.
2. UPF has the potential to greatly diminish the influence of our own contemporary cultural context and biases when it comes to the preservation of digital works.
The significance of UPF is that it is a recognition that there is a new paradigm when it comes to the preservation of digital cultural material. The "old paradigm" for the traditional materials that comes into archives, like books, letters, film etc is that cultural material is made of matter and, as such, it decays -- both through physical and chemical means. At some point with traditional materials we as conservators intervene, by upgrading the storage environment or actually doing a treatment, say to de-acidify a book or to mend the tears in a stack of letters. These treatments carry a cost for the object. The price the object pays is that the piece is altered, its no longer the same object. As conservators, we are trained in techniques that are designed to minimize and document these alterations, though of course what seems an appropriate change today can, with time, appear heavy handed or even ruinous.
Without a viable UPF, we are forced to apply this "old treatment" paradigm to digital materials. I would argue that its an old paradigm approach when we take digital cultural material made using one a "vintage" platform or file format and move it to the next viable combination of platform and file format. When that new combination fails due to impending obsolescence, we are forced to do it again and again hoping that these treatments somehow do not completely corrupt the "essence" and that the "preserved" piece (which is actually a new piece based on a lost original) somehow captures both the visual information and enough of the cultural, contextual information to make all the effort worthwhile. As I understand the promise of UPF, we are relieved of this burden since, as a neutral container, UPF provides a non-interventive means of maintaining the viability of the digital "essence" through time.
You really have to excuse this "new paradigm" & "old paradigm" stuff, I am as sick of hearing this language as your are. However, this "new paradigm" view is out there in a big way when applied to information and digital culture. For example, there is a series of essays, jointly published by the Association of American Universities and the Council on Library and Information Resources, called the "Mirage of Continuity: Managing Academic Information Resources in the 21rst C." This publication is a series of essays around the theme that when faced with "digital media, the conceptions of "Library" (and I suppose "Archive" and possibly "Museum") no longer serve institutions and may even be debilitating them." This is what I am taking about -- the old paradigm, the old modalities really may not apply when the cultural material held by an institution is digital.
For some, part of this new paradigm might mean migration and "Migration" applied to digital content is a term we are becoming accustomed to.
From my understanding, migration is the process by which digital content is ported from one viable format to the next viable format in an attempt to stay ahead of hardware and software obsolescence. Increasingly, migration based on taking content from one file format and saving it to another, more contemporary, file format is revealing itself as a potentially flawed approach:
Taking the first bulleted point: If you believe digital content has intrinsic value as cultural material, then a migration strategy based on creating a new digital surrogate is not a preservation technique since really, a new digital object is created based on the original. Furthermore, the new digital surrogate is often significantly altered as compared to the original.
One of the weaknesses of migration is that file formats come in all sorts of varieties and flavors and selecting the best possible migration path for one type of media might not work all that predictably for another. A quick and simplistic example of this point:
This is a screen grab of the Debabelizer www site. Debabelizer is a market leading file format conversion utility for media producers.
This product is especially useful for creators of digital media content that want to integrate all sorts of diverse media, from a wide range of sources, and repackage it using a more limited range of file formats.
Here is an example, a testimonial from the producers of the X-files CD ROM.
Debabelizer was used for "file format conversion, compression " etc… and "saving graphics to different formats"
Implied somewhere here is that the conversion process was done in batches, which is a real time saver if you are moving lots of images or Quicktime clips, whatever, into say AVI or MJPEG, whatever.
Now, I am not knocking this product, it is a fine, highly evaluated tool that get used by lots of people working under amazing deadline pressure everyday.
However, I just want to make the point that this conversion process we are talking about is fraught with peril and not without risk.
This is another screen grab of the Debabelized FAQ page:
Notice the second bulleted point under "File Formats." If you encounter this type of problem during a conversion you would no be too happy, but at least here you would be spared the absolute disaster implied by the second bullet from the bottom: "opening certain files results in 1 pixel x 1 pixel image."
So back to the first bulleted point: Not only is migration from one file format to another a poor strategy if you consider the source material to have intrinsic value, but it can be a somewhat unreliable course since seamless content conversion from one format to another can introduce artifacts and many other unintended variables that can significantly alter the "look" of a piece.
Looking at the second bulleted point: Setting aside the concerns for the intrinsic value of the original digital content, migration simply presents too much work to be sustainable over the long haul. There is too much to do now and and situation will only compound in the future. The current migration situation is like moving your household by sending a moving van on a separate trip for all the forks, then all the rugs, then all the lamps etc. Eventually you might get moved, but it would be painful and take a long, long time.
As an archivists you are in the same situation when it comes to migration. You need to set up and test individual migration paths for all your Quciktimes, all your d3’s, all your d1’s, all you MPEGs etc… which would be painful enough if the situation was static, but it isn’t and new media types keep coming in the door, needing their own optimized solution.
Just last week I got the annual report from the The Council on Library and Information Resources. This past year they commissioned a report from Jeff Rothenberg, a computer scientist at RAND "to document and assess existing models of digital archiving." The report argues that "migration strategies are simply too labor intensive to be viewed as a reliable preservation treatment, especially in an era of such dynamic change when standardization, which is critical for migration, is not feasible."
So, migration may prove to have serious disadvantages. Over the past year, the Council on Libraries and Information Resources commissioned Cornell University to formally provide a risk assessment of and comment on migration as a preservation strategy. I am anxious to examine that report in light of the problems we are talking about.
To summarize the problems: migrating content from one format to another significantly alters the original, creating a digital surrogate that may or may not resemble the original and clearly separates the content from its "context." Also ,the present situation of a "tailored" migration path for various types of content based on file format may well be unsustainable.
Given these disadvantages, implementation of a UPF strategy would go a long way towards eliminating the need for migration based of file format obsolescence.
First, once digital media is "enveloped" by a UPF container, the original attributes of the file format are preserved and documented.
Second, no further migration (based on file format) needs to occur. This presents a huge efficiency gain, eliminating the risks and labor continually associated with plotting "best-case" migration strategies for a broad range of file formats.
To conclude: I really have focused on how I think UPF would work from the standpoint of best preserving the original attributes of digital cultural material, essentially the first part of this equation.
The real point here is that preserving the integrity of the original digital file format or "essence" through the use of UPF (something we just talked about at length) is only part of the picture from the practical standpoint of archives. We all know that these bits have to be stored and that storage media can be unreliable (in the "old paradigm" sense of deteriorating through mechanical and chemical processes and through constant hardware obsolescence issues). For fine art pieces we also need to consider what playback equipment or other hardware is essential to the integrity of a piece, among many other concerns.
So even it all of our "essence" (our digital cultural material) is preserved through UPF implementation, there are still tremendous challenges posed by preservation of digital cultural material. However, by addressing a significant part of the problem puzzle, implementation of UPF could mean real progress toward the preservation digital cultural material.