Interview

Preserving History in Multimedia: An Interview with Stanford's Michael Keller

• By Linda L. Briggs
• 07/09/08

Michael Keller is well known as the innovative university librarian and director of academic information resources at Stanford University, as well as publisher of the Stanford University Press and HighWire Press.

Keller is also involved in another venture: a worldwide effort to digitally preserve vast amounts of material from history, both aging paper documents and very recent digital content. We talked with Keller about the effort, which is being spearheaded by both Stanford and Sun Microsystems through a group called the Sun Preservation and Archiving Special Interest Group (PASIG).

Campus Technology: Why is it so critical to get this preservation effort going now?

Michael Keller: The truth is that we've already lost a whole bunch of stuff. The loss started as technologies began to evolve....

The rate of change has increased, and therefore we have material around today that's really, really hard to read. If you try today to read a WordPerfect document you created 15 years ago, you'd have a very hard time doing it. That's just an illustration of the problem. We have a whole bunch of material that's no longer readable [or] that's becoming unreadable, and, if we don't take care of that, we're going to be missing documentation of our world--not a good thing.

CT: So part of it is a format problem. Is part of the problem that any media simply fades over time?

Keller: There's a bit of that, but mostly it's formatting. It is true that certain carriers especially tend to dissolve or change their characteristics [over time].... For instance, magnetic tape has a physical carrier which is plastic and a magnetic carrier which is some kind of metallic oxide. Those two expand and contract at different rates. If you leave them alone and you don't rewind the tapes once a year or so, eventually, the magnetic carrier and the physical carrier separate. When that happens, you may not be able to play [the tape] at all, or, if you do run it through the readers, you may discover that it's the last time you can read them.

CT: You can never put that back together again.

Keller: Right. There also are data format issues, application issues, and operating issues. Each of those has rates of change and those rates of change have to be accounted for in the digital archive.

CT: So there are quite a few challenges there.

Keller: Yes, but fortunately we have a number of techniques we're testing now. And we have some corporations, especially Sun, who are very much in the arena with us and developing some really affordable and terrific technologies.

CT: As a librarian, how do you decide what to keep? How is that decision being made?

Keller: First of all, understand that each institution has different targets, though some of them in the academic world overlap considerably. But among all the cultural agencies--from museums, libraries, and archives--there's enough diversity so we cover a fair swath of what's there.

Mostly, [at Stanford] we select on the following basis. It's complicated and most of the selections are made by people that are very well advanced in whatever disciplines they're working in, often with Ph.D.s. I have 35 or 40 of these people that do this here right now, today.

First of all, we are intent upon documenting the history of Stanford itself in as much depth as we possibly can. Because Stanford's big, old, and very busy, there's a lot of that to deal with.

Secondly, we collect deeply for the practitioners that are here right now, today. So, if we have 45 individuals in the history department covering ... history, we address each one of them as a particular client. We treat all of their interests collectively and very deeply. That is, we have several who are interested in Islamic history, Middle Eastern history, Russian history, British history, American history in several different periods and so forth. We collect deeply in all those areas in order to support them in their research and their teaching....

Thirdly, we collect representatively on a global basis from what we used to call the book trade, and we now call the information trade. For instance, what may be published in Hindi in India with regard to sixteenth-century political history there is of interest here, even though it's India and we don't have a South Asian program. We will collect that book or those books because we are generally interested in medieval and renaissance history.

Also, we collect special collections material that match our interests and provide grist for the scholarly mills, but we also collect as cultural custodians. So as something becomes available--a corporate archive, a personal archive, or an odd thing that we know about ... because we have a special relationship--we try to acquire that material as well....

CT: Clearly, we're not talking just digital formats. We're talking about books, magazines, and so forth as well.

Keller: Yes, all kinds of things. The day I arrived [at Stanford in 1993] ... I said to everybody here, "We are no longer interested in the medium on which the information or the knowledge comes to us. We're interested in the knowledge and the information. Then we'll cope with the media."

The problem is, of course, that the digital media present us with special preservation and access issues that we must cope with....

CT: Clearly, you saw this coming back in 1993.

Keller: I actually saw it long before that, but I wasn't the boss here until September '93. I was at Yale before I came here.

CT: What percentage of what's out there do you think is being saved, not just at Stanford but everywhere? Is it just a tiny percentage?

Keller: Yes, it's fairly small, I'd say. Overall, I would say that places like Stanford collect maybe 3 percent. Places like the Library of Congress are [collecting more] in terms of percentage.

In certain domains--for instance, academic publications--we're very rich, probably more like 75 [percent] or 85 percent coverage. But in other domains, such as popular graphic novels from Japan, we only have a few representative samples. We don't have anybody working on that area just now.

CT: It depends on who is focusing where?

Keller: Yes, but in that case, we count on the Japanese libraries and national collecting agencies. We have a pretty good system of knowing what's where and either sending the stuff or representations of it, or sending the people that want to look at them, to where the material is.

CT: Looking at the other side of this, what about the access implications? How do you decide who can access what, including Web 2.0 material? Is that a big issue?

Keller: Yes, it is a huge issue. Digital rights management is huge.

We have several different frameworks in which to work. One is the framework of laws and regulations; libraries, archives, and museums are especially attentive to that. We are intended to provide access to the public, but that doesn't mean we can let the public walk away with something. In some cases, we have to make sure we get the right licenses, pay the right fees, and so forth to enable people to look at documentation. We also have to be careful about how these materials are used in classes or beyond classes. We don't permit wholesale copying. It's wrong.

On the other hand, there are some materials that are owned personally.... [With] personal papers and archives, we sign contracts with the owners. Although we won't take something unless it ultimately becomes accessible to our scholars and to the public, we sometimes take collections, parts of which become publicly accessible only after certain events occur or certain numbers of years have passed.

For instance, if we were to acquire the papers of a governor of California--and we actually have some governors' papers--most likely, the governor and his family would say, "Well, there're some things we don't want to be made public for 25 years," or "... until my children die," or something like that.

But ultimately, those materials can become grist for the scholarly mills.

CT: So it's complicated.

Keller: Yes it is. But once again, the same principles apply to physical items and digital items. We have to develop some new means of coping with the digital items. That's what we're working on. But the principles are well evolved.

CT: I wonder if you can give me an example of something that you've been able to preserve that you might not have, had you waited longer to begin this project.

Keller: A little while ago, thanks to support from a local family, we were able to acquire a collection of almost 400 antique maps of the colonial period of Africa, particularly Southern Africa. This collection was based in Johannesburg but there wasn't the local interest or maybe the local wherewithal to acquire and take care of this collection. But in acquiring it, we agreed to make these map images available to the world and especially to people in Southern Africa and South Africa.

So we digitized them, and they are accessible. But we need to be able to save those digital images so we don't have to redo the digitization, the scanning, because some of these maps are quite old.

That's an example of something that goes into the digital repository, where it will be maintained and as a result be accessible, regardless of the way you view it, over many, many, many generations.