Joan reported that Meg Bellinger had asked the committee to consider having a town meeting style forum on metadata to gather views and experiences from around the university. This was discussed further as the second agenda item.

As part of the start up of this committee we are having a number of reports by committee members on current activities, projects, and programs with a metadata component.

Today, Rebekah Irwin reported on metadata for the Beinecke Digital Library. The report and discussion were very useful.

Rebekah distributed 2 handouts on the Beinecke DL bibliographic record. One showed a blank record form with all the data elements. The other showed a completed record.

Rebekah summarized the Beinecke DL and noted that the Beinecke DL has 18K bibliographic records for 94K digital scans.  Clearly, 1 bibliographic record can be linked to more than 1 digital image.  The bibliographic record itself does not contain any reference to the corresponding digital image. This information is tracked in an online Access database maintained by Luna and in the SQL database behind the Beinecke DL. A separate table is used to link the appropriate digital images to their corresponding bibliographic record number.

The production rate in the scanning studio at Beinecke is 120 per day (60 with camera, 60 with flat bed). Cataloging must keep pace with scanning.  Cataloging work precedes the scan production work and is used partially to shape the scanning work flow. Images go to Luna for post-scanning work.

Curators select about 70% of the scanned material (e.g. exhibitions) and patron requests drive 20 to 30% (e.g. publication or classroom use.) Sometimes a patron request grows into a curatorial request. For example, a patron may ask for an image from a manuscript, and a curator may decide to have the whole manuscript scanned.  Beinecke has yet to formulate collection development policy for digital scanning and collections, but is, generally, moving from a system driven by patron requests to one driven by curatorial decisions.

Rebekah gave a demonstration of the public view of the Beinecke DL. Again, discussion was aided by many good questions.

Issues discussed included:

• How multi-image objects can be usefully presented to the user. Two views are possible: gallery show (several images seen at once) or slide show (one image seen at a time.)
• Type of image made by scan (400 dpi TIFF) and number and size of derivatives
• Should links from the Beinecke DL to other tools (like Orbis) go to the record itself or to the a higher level of the tool. With the former one loses context. With the latter one has an extra step.  This is similar to the issues with SFX links. What is best for users?
• Subject term access for images in Beinecke DL. Subject terms now in use, but had not been until recently. DL has no authority control module within it, but terms from controlled vocabularies can be used.
• Data clean up needed for DL.
• Cataloging about the _original_ object. Technical metadata about the digital image is in the database but not explicit in the bibliographic records.
• Rights information only at the highest level in DL and not in each record.
• The metadata input tool has good editing features that allow for quick production of record for items within a scanned collection (e.g. 1400 lantern slides of American Indians from McClintock papers.)
• How is the DL promoted? How is the DL incorporated into teaching, research, and publication? Critical issue. The DL is being publicized within Beinecke and some user circles, but DL is a silo that users may not know exists and thus cannot use.
• Beinecke DL not now doing TEI markup of texts that have been scanned.
• Harvesting metadata from DL for reuse elsewhere not yet thought through.

Matthew distributed a rough outline of issues that would need to be addressed to put on a useful forum on metadata.

Questions included: What is our objective? Who would attend? What would it be like? When? Where? What results or outcomes would be expected?

Brief discussion indicated that all these points need work. Further discussion was put off until the next meeting, but it was clear there is interest on the committee in doing this, that we had to think about the whole university (not just the library) not just in terms of who should come to the meeting or forum but what issues of metadata and use we need to consider.