Over the last two years, the Arquin Project team has worked on creating metadata for the slides in the collection. We decided early on to use Dublin Core for our metadata terms since it is one of the most common metadata schemas for web content and has been adopted by Omeka, the platform we use to make the collection publicly available. We started to familiarize ourselves with the metadata elements and decided to use the mandatory fields and a few others.
One of the additional fields that we included for the items is coverage. This field allows us to record the coordinates of the place in which the images were taken by Arquin during her trips. The other main fields that we have for each item are title, subject, and description. The data for these fields has largely come from a printed catalog of the collection produced, we believe, in the years after its acquisition. A large percentage of slides include some description and geographic location and this has facilitated the process of creating metadata. We use Getty
Our students working as metadata specialists for the project have copied information from the OCR-scanned catalog pages to fill the title and description fields. For instances where there is no catalog data, we have resorted to looking directly in the slide mounts for Arquin’s handwritten annotations and have based the descriptions on those. During this process, we encountered spelling mistakes in the original materials and made an effort to correct them in our metadata spreadsheet. For the subjects, we have used Getty’s Art & Architecture Thesaurus® Online to keep our subjects within the controlled vocabulary.
In support of the process to create metadata we have created an internal document that we called “Guidelines for Metadata Input, Correction, and Translation”. This document has helped us determine the steps involved in the process of creating the metadata for both the cataloged and uncatalogued slides. The document also includes notes on our process of making corrections to metadata. This part of the process has been supported by an information sciences consultant. By contrasting Getty’s subjects with our subjects we were able to identify subjects that didn’t match the controlled vocabulary or that had a spelling mistake. This became an essential part of the process as we moved towards establishing our guidelines for metadata translation. The next blog post will cover the details of how we managed the translation of certain metadata fields.
Leave a Reply