Originally posted on 2014-04-01
Reviewed by Erin Lawrimore, The University of North Carolina at Greensboro [PDF Full Text]
Developed as part of the One Week | One Tool open-source software development institute during the summer of 2013, Serendip-o-matic is billed as a “serendipity engine.” A user inputs text (anything from lecture notes to song lyrics to a full bibliography), presses the “Make Some Magic!” button, and the “Serendhippo” (a cartoon hippopotamus that stands as the Serendip-o-matic mascot) appears as results based on keywords selected by the tool’s algorithm are built. These results include materials from the vast resources of the Digital Public Library of America (DPLA), Europeana, Trove, and Flickr Commons. Thumbnails are displayed with a title viewable as scroll-over text, allowing a quick visual browse of the results of the search. A user interested in more information about a single record can click the thumbnail and be taken to the website that hosts the full digitized record and its accompanying metadata.
Serendip-o-matic’s search results are “designed mostly for inspiration.” They are “not meant to be exhaustive but rather suggestive.” Brian Croxall, co-project manager for the tool, explained in a post on the Chronicle of Higher Education’s ProfHacker blog that “the goal of Serendip-o-matic is to create opportunities for surprise . . . something akin to finding that unexpected book in the stacks or that item you didn’t know was in the archives.” In other words, the tool is not a traditional, ranked search engine intended to unite the various digital repositories. Instead, it serves as a source for learning about and making connections across repository lines.
Serendip-o-matic has numerous potential uses for archivists and archival researchers. For archivists who are looking to work across institutional boundaries to collaborate on digital collections, exhibits, or other projects, Serendip-o-matic can reveal other institutions with similar holdings. For instance, a search using a finding aid’s historical note can uncover digitized records related to key people and places. Searching using the text of an upcoming presentation or blog post can return open-access images that can be freely used as illustrations. Additionally, Serendip-o-matic can link to a Zotero account, allowing a researcher to search for materials related to those already captured in the citation management tool.
I conducted a search using text from a web exhibit focused on women’s basketball at my institution in the early 1900s. This search returned multiple items related to the creation of basketball by Dr. James Naismith as well as results related to other school’s women’s basketball teams. Included were digitized images of women’s basketball competitions from an archives located only hours away from my own. With this information, my research could easily expand from the local to the state to the national or international levels. This benefits archivists as well as archival researchers who are hoping to uncover related materials across the globe.
The primary weakness of Serendip-o-matic is not related to the tool itself but to the content it aggregates. DPLA, Europeana, Trove, and Flickr Commons aggregate records from a number of content sources, resulting in a wide range of resources that effectively meet Serendip-o-matic’s aim of providing a broad-reaching, serendipitous search experience. The metadata associated with the content across these aggregate platforms, however, varies greatly. Because Serendip-o-matic is meant to be “suggestive” rather than “exhaustive,” this does not strongly affect the value of the results returned. The user will still return a multitude of records associated with the text inputted. The results display, however, is limited if the associated metadata is absent. Typically, Serendip-o-matic results are displayed as thumbnail images with a brief title viewable when a user scrolls over the image. If the aggregating site has no thumbnail associated with the returned record, the Serendhippo appears as a placeholder. If no title is associated, the scroll over text states “Title Not Known.” Users can still click on the thumbnail (or Serendhippo placeholder) and be taken to the full-digitized record. But returns such as these can interrupt the visual browsing of the results that can be so important in this type of serendipitous discovery environment.
In spite of possible issues related browsing displayed results, Serendip-o-matic fulfills its mission of being a “serendipity engine.” Users – archivists, researchers, and others hoping to find interesting primary sources from around the world – can uncover surprises in each set of search results. Additionally, because Serendip-o-matic is open source with the code freely available on GitHub, those who wish to modify or extend the tool are able to do so. The overall result is a search tool that can spark new ideas and projects, and help identify and build connections that span institutional boundaries.
 Brian Croxall, “One Week | One Tool: Introducing Serendip-o-matic,” ProfHacker, August 5, 2013. http://chronicle.com/blogs/profhacker/one-week-one-tool-introducing-serendip-o-matic/51449