This week, as my class explores how to effectively use metadata on the history web, I examined the following two websites that present oral histories documenting African American history to evaluate their use of metadata.
Behind the Veil
Behind the Veil: Documenting African American Life in the Jim Crow South is a website I first learned about last semester in my oral history course. The site allows you to browse all 410 interviews if you desire to do so, but also helpfully categorizes the interviews based on their metadata to make them more useful to users. Users can type in a keyword search browsing this collection or several related collections, or they can search tags categorized by state, interviewee, birth state, gender, occupation, interviewer, and series. The metadata page for each interview provides useful entries for each of these categories, as well as information about the interviewee’s occupation and the like. A brief bio of the interviewer and interviewee, or maybe a short abstract of the interview would enhance the navigability of these interviews for researchers, but the subject category provides some assistance. Although users cannot necessarily contribute to the available metadata, there is a comment section beneath the metadata of each interview where users may contribute their reactions, insights, observations, recommendations, or related research, creating the opportunity for a richer wealth and repository of information not only of the interviews, but of how people today have used, valued, or responded to these interviews. Users may also submit questions to the site’s creators and research team.
Voices from the Days of Slavery
Like the Behind the Veil project, Voices from the Days of Slavery, an oral history collection on the Library of Congress website, contains ample metadata, but is not necessarily as search and research-friendly as the material on the Behind the Veil site. Like Behind the Veil, one may browse the collection through the Audio Interviews, Song Titles, Names, Subjects, and Places categories, but the available metadata is buried beneath several links, difficult to navigate, and provides very little contextual information. The metadata contains little more than the title of the interview (listing who was interviewed) and the repository location for the physical audio tape. It appears that some valuable metadata is missing from the site. Some related pages provide contextual information about some of the interviewers and interviewees, but not all of them, leading to a skewed and incomplete knowledge of the context of the interviews and interviewer-interviewee dynamics. For instance, biographical information is available for all of the (mostly white) interviewers, but only for a handful of the interviewees. Nor is this material directly linked to the specific interview to which it pertains.
The Library of Congress site has a standardized system for its use of metadata and does not appear to allow user contributions. This may be a detriment to the project, but not necessarily so. Too much user-contributed content may further hinder smooth searchability and research on the site, rather than aiding it.
Sometimes, however, metadata can be distracting if one is trying to use the web to present an online exhibit or some other theme-centered history media. Users may get lost in too-specific information about an image or item they don’t need when they just want an overview of a historical topic. Maggie, Fazila, and my upcoming Evanston Women and the 19th exhibit endeavors to make metadata available but keep it from distracting from the theme, storyline, and contextual content of the exhibit by interpreting our collection material using an interactive chronological timeline divided by local, state, and national contexts. Visitors to the site can click on each entry to the timeline to view more information about an event, which will connect thematically to other entries on the timeline. Timeline entries will also contain images, audio, video, and other primary sources that the user can click on to obtain more information. A “Read More” button on the entry will direct users to our Evanston and the 19th Omeka site, which will provide metadata for each item in the collection for those interested in more detailed information or research material. Users can browse material in the site more effectively using the tags and record type links, as well as the keyword search, which enables users to narrow their search more specifically.
Nice review. I share your concern that while it is critical for online collections and databases to make metadata prominently available, the same metadata can be distracting if displayed too prominently as part of a thematic exhibit. I touched on that in my blog post as well. One challenge that neither of us addressed, though, is how exhibits that only provide metadata by linking to it on a separate database page might limit the discoverability of the exhibit itself. People who search the web for items in the exhibit might easily find the database record full of metadata keywords, but fail to find the exhibit section that interprets the object as part of a broader theme or story.
Perhaps one way to address this would be to include a metadata element in each item’s database record showing what exhibits the item appears in, with hyperlinks going there.
Another possibility might be to include collection metadata on exhibit pages in a format that is machine-readable for search engines but that is not immediately displayed to human viewers. The problem with this is that search engines typically prioritize human-visible content, since hidden metadata can so often be dishonest or misleading. Making this work would also require exhibit builders and search engines to agree on supported standards for encoding hidden collection metadata.
LikeLiked by 1 person
Excellent points, Josh. Knowing much less about website design and web content than you do, I had trouble conceiving of as creative alternatives as you did. How might one new to web development with limited time and financial means be able to go about your recommendations? The hyperlinks option certainly appears manageable, but the machine-readable metadata sounds more challenging. Would users still have access to the machine-readable data if they wished to view it for research?
LikeLike
Good questions. I agree the hyperlinks option is more straightforward, and when using Dublin Core in Omeka, for instance, it would probably be perfectly appropriate to include such links in the “relation” attribute that is already provided.
The machine readable metadata is more of an imaginative solution than a practical one at this point. My thinking was that this would just duplicate the visible metadata on the item database page as hidden metadata in the exhibit for search engines. Human users could still access the visible metadata by following a link to the database. They could also find the machine readable code itself by viewing the HTML source of the exhibit page, which most browsers allow. Actually implementing something like this is another matter, though. Google does offer guidelines for embedding machine readable “structured data” in web pages, using existing RDF or JSON-LD encoding standards — you can find their guide to this subject at https://developers.google.com/structured-data/ — but the uses they have embraced are targeted towards business , social media, and advertising rather than historical, academic, and cultural repositories. I think a fair bit of development would be necessary before this became an accessible strategy for online exhibit makers. Given that the other solution already has a practical implementation, this more complex one might not be worth the effort.
LikeLiked by 1 person