Summary of MDG session, 1-29-08

* Written by Jenn Riley *

The second meeting of the Metadata Discussion Group was a lively session, with about 50 people in attendance. We’re glad to have a new exit installed in the staff lounge so that we no longer have to count heads to keep under room capacity!

The article for discussion in the January session was:

Elings, Mary W. and Günter Waibel. “Metadata for All: Descriptive Standards and Metadata Sharing across Libraries, Archives and Museums.” First Monday 12, no. 3 (March 2007). http://www.firstmonday.org/issues/issue12_3/elings/index.html

The discussion began surrounding to what degree the article successfully described the different perspectives of the library, archives, and museum communities. Some in the group thought it was a good introduction, but didn’t make clear possible differences in broad vs. narrow scope or unique vs. non-unique materials. One participant noted that these institutions all have collection management needs in common.

Overall, the group felt that choosing standards based on material type rather than institution type made a great deal of sense. Visual materials were mentioned especially as needing a different approach than textual materials, and this participant postulated that this new way of thinking has been influenced heavily by the visual resources community, where content and structure standards developed side by side and influenced each other. Many questions about how to put this approach into practice were raised, however. Are we breaking user expectations by using different models for different types of materials? Some of our current categorizations of materials are hard to separate out by type – government documents, for example, are a mixture of archival-style things, visual materials, and texts, and a complete collection of the photographic work of one person would benefit from the context provided by archival description but would also benefit from in-depth item-level indexing.

The group discussed for a time why things developed they way they have. One participant noted that the article (politely!) casts blame and says we insist on using the wrong tools because we’re used to them. There’s a reason we use the tools we have, because we have financial, administrative pressures to produce more. We care about interoperability, but we can’t afford it. Implementing a major shift in how we approach things has enormous financial implications. No good models for institutions making this shift on a large scale, and our technological tools haven’t caught up to this new way of thinking either. We need to “expand our personal toolkits.”

The group spent a significant amount of time discussing issues of efficiency and streamlining the descriptive process. The Greene/Meissner report cited by the article was mentioned, as was RLG Programs’ recent report: Shifting Gears: Gearing Up to Get Into the Flow. The disconnect between collection- or file-level description and item-level digitization was noted, and seemed problematic. Questions of efficiency led us to discuss user-contributed metadata, especially as seen in the recent LC experiment with Flickr. Participants felt this approach could shift some burden from us to our users, and increases exposure for our collections. But we still need to process collections, and do a great deal of work with them. We need to experiment – things will keep changing. Questions about the need for oversight of user-driven data were raised, with overall acknowledgment of the problem but no concrete solutions.

Participants raised some specific questions with regards to the article:

  • A version of the chart in the conclusion adding visual resources as another category was distributed at a conference last summer. Is this simply another category or does it change the argument somewhat?
  • Could the chart in the conclusion include MARC as data structure for each community?
  • Where does Dublin Core fit in all of this?

The session ended on a somewhat philosophical note, with participants commenting that the shift in thinking proposed by this article reflects changes in society in general – more collaboration, themes emerging between groups are happening. The social definition of knowledge is changing. The group closed by noting interesting content on the web and remembering that the “interesting stuff” is why we do all of this in the first place.

Our next session will be Tuesday, February 26. The article for discussion for that session will be distributed by February 12, Please send ideas for topics for future session to the MDG listserv, and feel free to use the listserv for discussion between in-person sessions.

Summary of MDG session, 11-27-07

* Written by Jenn Riley *

The first meeting of the IUB Metadata Discussion Group seems to have been an unqualified success. Although we held the session in a room with the largest seating capacity in the Wells Library, we still had to turn some people away. My apologies go out to those who wanted to attend but were unable to because of space. By our next session, a second door should be installed in the room which will raise its legal maximum capacity. Thank you to all who attended, and who tried to.

The article for discussion this month was:

Gilliand, Anne J. “Setting the Stage.” In Introduction to Metadata: Pathways to Digital Information, ed. Murtha Baca. Online edition, version 2.1. Available from http://www.getty.edu/research/conducting_research/standards/intrometadata/setting.pdf

The group used the general principles outlined in the article to discuss the role of metadata in libraries and their technical services departments. Participants appreciated the breadth and high-level focus of the article, but expressed an interest in balancing this approach with more practical approaches in future meetings. The difficulty of describing the concept of “metadata” in any succinct way was noted by participants.

Two features of the article were brought out in discussion: the thought of metadata as something that grows and changes over time, and the fact that “lay” metadata is important in addition to “expert” metadata.

Regarding the continued accrual of metadata over the lifecycle of an object, the group discussed the potential effects on copy cataloging of this need, noted that WorldCat Local could play a part in this, and postulated that one of the roles of a technical services department could be the adding to of metadata records over time.

The concepts of “Lay” vs. “Expert” metadata, not surprisingly, generated a good deal of discussion. No participant voiced the sometimes-heard opinion that metadata from lay sources such as users, publishers, etc. (including user reviews, sales data, tagging of images, etc.) had no place in the library environment, although several individuals cautioned that the metadata we maintain must support effective retrieval and that more uncontrolled metadata could threaten that goal. One participant voiced an opinion that one role of libraries is to supplement lay metadata with expert metadata, to help ensure authority, a sentiment that seemed to have general agreement.

From this point, the discussions turned to the role of systems in providing services based on metadata. Participants felt that our systems needed to handle both factual, structured data like ISBNs and more fluid, organic, unstructured data like that our users can provide. It was noted that to provide high-quality services on these different types of metadata, our systems need to have *more* structure on the back end, rather than less. While the discussion didn’t delve very far into specific metadata formats, there was a general sense that the data being recorded was more important than the format in which it was stored. One participant summarized this view as “I don’t need to have MARC, I need to have the specificity of MARC.” The need for different approaches for different types of materials was raised, which led to a request for future MDG sessions to study in more depth these different approaches, the standards that emerge from them, and the communities behind them, and to discuss whether there is more these communities can do to work together. Participants also expressed interest in system design issues, allowing complex linking of records but still allowing them to make sense out of context.

Throughout the discussion, possible roles for technical services staff in the metadata environment emerged. Most ides centered around creating and maintaining descriptive metadata, although a need was expressed for all involved in metadata creation to know about all types of metadata being created and how it is used. Possible roles for technical services staff included:

  • Recording relationships between information objects that are not possible to generate automatically. (For one initiative designed to help automatic recognition of relationships between objects, see http://www.openarchives.org/ore/)
  • Authority control, to allow more powerful discovery mechanisms
  • “Expert” metadata to supplement that from other sources
  • Describing hidden collections with no or inadequate existing descriptive metadata
  • Describing Web sites intended for archiving
  • Describing objects deposited into IU ScholarWorks
  • Targeted projects to enhance older metadata
  • Provide value-added content
  • Managing groups of records
  • Providing acquisition information to fund managers

We had a lively discussion, with many points of view raised. Our next session will be Tuesday, January 29. The article for discussion for that session will be distributed by January 15. Please send ideas for topics for future session to the MDG listserv, and feel free to use the listserv for discussion between in-person sessions.