Lighting the Way
Vicky McCargar
Los Angeles Times

Archiving Update
Monday, June 7, 2004
12:15-1:30 pm
Special Libraries Association 2004 Annual Conference News Division

The more we digitize for access, the more we put at risk because digital media is very fragile. The more characters we put on something, the shorter its lifespan seems to be. Things are still advancing and changing. Some archivists think the data from the 1990s are already gone because of access issues: the speed material disintegrates, the lack of metadata and proper indexing, etc.

Is serving future historians our mandate? It's difficult to know what's important, which puts us in the position of keeping everything. The value of material changes after 90 days. Digitization is about access: reusing material, outside sales, and sharing.

News libraries have similarities with other cultural institutions because things in our collections are similar and the work that needs to be done to preserve them is very similar. There aren't very many standards for what we do, but our data is growing more complex all the time. Media that is born digital is the most fragile of all media. Digitization is not preservation, especially as things stand now.

Vicky polled the audience: How many people still have the original graphics from the 1991 Gulf War? No one raised their hands. Vicky says the LA Times still has them because she's been keeping an eye on them for a few years.

Large institutions are grappling with these issues, like the Library of Congress, OCLC, NARA, universities, and many others. Countries are also working on them. So far, issues include migration, emulation, encapsulation, XML decomposition, normalization, and archaeology. Metadata schemas are even working on these issues. The amount of data librarians should keep to make sure files will be accessible in the future is enormous. Third parities are registering types of software in a format registry. Third-party repositories are probably going to be the next big business.

The Open Archive Information System (OAIS) is setting some standards.

Be careful what you call systems: some managers balk at the word "archive," but many are interested in "digital asset management systems." The Los Angeles Times is working toward their own standards and schema. News libraries often focus on a printed product, but archiving begins when the reporter sits down to write the article. Creativity among users isn't necessarily supportive of an archive. System flexibility and software is important. "Keep" "everything" "forever" (in quotes to emphasize the words aren't really conducive to keeping everything forever).

Budgeting for preservation is expensive. It's important to think about that within the research budget. Hard copy photos are less expensive.

We need to strike a balance between the latest and the greatest software versus stable platforms. It's also important to balance between trends, like how some graphic artists are now saving their work in the basic files, which usually don't last more than two years.

What the future looks like: lots of long-term costs and experiments. D-Space from MIT seems promising. Maybe there will be standards. We'll need to work with vendors. We need metadata and migration strategies at the beginning of the workflow.

Vicky is trying to get a group of news librarians together to examine metadata issues within the news librarianship realm.

Question: something about metadata
Answer: Look for information about MODS on the Web. Metadata standards are still indevelopment.
Rob: WGBH in Boston and others are developing industry-standard metadata for broadcasters.

Question: What do we do with our digital media?
Answer: Plan for migration from the very beginning. Anticipate that migration is necessary. If you have anything on CDROMs, get it off. They only last 5-7 years. Look at what you have and decide what you need to do with it. Proprietary formats are difficult to keep alive. People are working on developing an archival .pdf format now.

Question: about versions of CDs and if certain brands are better than others
Answer: Standards could help manufacturers make something that will last longer, but part of the problem is changing software and computers.
Rande: a trend among their clients is to get rid of CDs and CD jukeboxes and move to RAIDs and other nearline storage.
Rob: Don't trust a CD or DVD for more than 3 to 5 years. NPR uses an error detector to check if CDs are going bad.
Many newer CDs only copy at high speeds. NPR copies in real time. They had to place a special order because high speed CDs won't work for them. CDRs use a laser to write. Light can damage the CDs.

Vicky suggests using metadata to mark files that need to be migrated, especially those that'll need to be migrated at a certain time. Technical metadata includes metadata about the technical aspect of the file, including when it will need to be migrated.

Question: about tying preservation planning to the production cycle
Answer: Vicky recommends tying it all together. It's something we need to plan from the very beginning of production.

Rande: Rights management: some archiving systems allow for rights tracking and rights management. Someone asked about that earlier.

Special Libraries Association 2004 Annual Conference
SLA News Division Home

Last Updated: 06/07/04