Sharing our descriptions

In class we discussed how photographs should be described.  Our talks revolved around the act of creating a good record and the steps that one should take to produce a complete record.  One aspect of creating a good record that was lightly touched upon was the idea of sharing the photographic description.  Sharing photographic description or any metadata is an important activity that we need to engage in as our world becomes increasingly intertwined. 

This idea started to interest me after examining the two implementations of LIFE’s photo archive.  I believe that if Google could harvest the additional metadata present in the Getty it would be a much more useful site, given it corrects the issues with how the metadata is linked to the photographs.  There is no doubt that people would prefer Google’s version to that of the Getty.  This is in part because Google’s version is simplistic, it does not have a watermark across the middle of the image, and more importantly it is familiar to searcher.  When conducting keyword searching, having more metadata would be crucial to a user being able to find what they are searching.

In writing a description for photographs, three objectives come to mind.  The description should give the ability to identify the photograph, then select it, and finally be able to obtain the image.  In conjunction with those objectives the qualities that contribute to the creation of a good record are completeness, accurately stating the aboutness of an image, describing the work type, describing the subject, consistency among records, and favoring the name of the creator.  Without quality information it would be pointless to share those records.  The three objectives are the pivotal in making photographs discoverable.  I say this because they help keep the cataloguer’s mind on the user, so as long as each characteristic is described with those objectives in mind quality records can be achieved.  The main reason we describe a photograph is to make it accessible and without those objectives being at the forefront of the task accessibility may fall to other concerns. 

This brings us to our next issue.  In order to make a photograph accessible, we must provide good metadata and in order to make the collection accessible to a large audience we need to share our data.  The problem is that when sharing our descriptions the metadata being shared is not always high quality.  One reason is because of the limitations of how data can be cross-walked.  The standards being used to create the record needs to be mapped properly to the standards beings used by the destination system.  The problem here is that there is no definitive cross walking map.  In addition there are so many standards around, that the number of mappings needed would be even greater than the number of standards.  Another problem is that there will always be some information that is loss during the crosswalk.  This makes me wonder if it would be better to direct users to the other institution that has done the cataloging of the materials rather than making your own in the sake of creating a unified user experience.

The “big fish” example brings up another problem that descriptive metadata faces, that of the quality of metadata being created.  There are so many descriptions available, how do we know who we can trust to have the authoritative version?  We could say an institution that does this for a living should be the authority.  However when looking at the Getty’s implementation of the LIFE archive, the description appears to be the caption of the photograph and we do not know who created the caption, most likely a photographer and not a cataloger.  A second example is one that was shown in class from the National Portrait Gallery, that stated the focus of the photograph is Laurence Kerr Olivier, and does not mention the theater that he was seated in which appears to be the large focus of the image.  If we cannot trust reputable institutions, then who can we?  What if we instead made a composite of all descriptions using the most frequent terms?  The idea of an aggregate description seems very appealing.  However it stills falls under the problem of the quality of the data being aggregated.  If you have a lot of bad descriptions and you put them all together, you end up with a consensus of bad descriptions.


There are a lot of factors that cross my mind when cataloging.  I ask myself the following questions.  Who are my users?  What are their needs?  How important is sharing the data?  Is having good interoperability important to my user group?  Will sharing the most basic metadata be enough for the discovery of my resources?  After I ask myself these questions I realized what is important and usually interoperability becomes secondary to the needs of the institution and their users.  It is valuable for us to be able to share our data, however before we can successfully do so we must address the issue of creating quality records.  The mechanism in which we share the data will come naturally if we create good data.
 
Reference
Shatford Layne, S. "Some Issues in the Indexing of Images." Journal of the American Society for Information Science 45, no. 8 (1994): 583-8.