Thursday, April 8, 2010

Ah Metadata...

Jason Roy, at the Digital Collections Unit/Digital Library Development Lab, University of Minnesota had a word of encouragement for all those archivists out there while speaking at the VRA conference this year.  "You don't have to catalog on the item level." 
One of his archives received funding to scan an enormous collection for which they just had finding aids. For those not up on archival cataloging, it is customary to describe a collection by elements within a box, possibly a folder. So, you might have a folder of letters to Mr. Deere over a period of time. Alternatively, you might have a box of the Deere family memorabilia, estimated to cover 20 years of their lives.  The archivist will browse through said folder or box, looking for the item he desires.  The finding aid will give him a general idea where to best look.
Most digital collections have insisted that archivist must now throw some metadata at each item as it is converted to a digital format, so that an item is retrievable. 
Well, Jason, upon being confronted with the daunting task of converting this large collection to a digital format in the conventional manner, said no. Rather, they would place digital files in the appropriately labeled folder according to the box / folder information that existed already and would published a finding aid to direct the searcher to the appropriate "holder,"  which the searcher could then browse looking for the desired item.  No promises that it was there.  This is not ideal, but it is as good as the existing condition.  When you also throw in the technical ability to tag digital files with further information as researchers retrieve material, it means it will be a growing evolving catalog.  In addition, if you then embed the descriptive metadata that you have in each file, the digital file will always be identifiable without being bloated.
This approach would free up so many of our archives, if more special collections could accept it.  What do you think?

1 comment:

  1. Trudy,

    I think that Jason Roy's approach makes sense particularly in the archives or special collections arena. His approach is both pragmatic and incremental; it allows for user contributed enhancements through social tagging. Of course, professional staff will always be able to add additional information as well.

    While I believe that words provide the key to image discovery, fewer words are better than none at all. Words can always be added to or subtracted from a record or an aggregate of records in an organic process.

    On the other hand, this approach works only with certain types of collections particularly those that are primarily important for research and discovery. Instructional image collections function in a different arena; I suspect, that item level cataloging for these types of collections is much more import because users are searching for a specific image no matter where it might be located. For instance, an art historian is looking for a digital image of particular Leonardo da Vinci located in a particular museum. No other image will do, and it doesn't matter where that image file resides. But even here, collection level records might well suffice; e.g., all of the images in "this" collection pertaining to Chartes Cathedral or the L'opera in Paris.

    I think that the real point here is that cataloging and descriptive work can be done incrementally as time and resources permit. Furthermore, the effort can and should be shared between catalogers and end users. Image records are often neither static nor complete; they will change over time. Less is better than nothing at all.