In our own digital an(ti)thology, student-authors utilized a wide range of content types and presentation options for their recovery projects. The primary media featured in our Scalar book consists of digitized selections from our university's Archives & Special Collections, including works by Eliza Bradley, Frances Power Cobbe, Dinah Mulock Craik, Lola Montez, Phillis Wheatley, Sarah Winnemucca, Mary Wollstonecraft, and an anonymously authored women's recipe book. Students selected excerpts for digitization, as many of the texts were not previously digitized locally by our Archives and Special Collections staff. Though digital versions were available elsewhere, it was important to us to engage the local specificity of our own holdings – an example of the local and global interface in digital archives (Graban et al., 2015). To prepare the selections for our production, the archival staff took photos using a digital single-lens reflex (DSLR) camera in Adobe Lightroom. The files were converted to TIFF files, edited for light, and saved for Scalar as JPG files. We then hosted these files on our institutional instance of Omeka, which we linked to our Scalar book. 

This solution to the hosting issue did present further problems at first. Specifically, we found that the data output by our institutionally-hosted instances of Omeka were not compatible with Scalar, as our servers did not allow the data to be scrubbed and embedded in an external website.  This setting was controlled with a header set at the web server level (Content-Security-Policy for frame-ancestors).  This header, though, could be overridden in a .htaccess file within our site, which gave the option to allow the framing of certain content on certain sites. If you have control over the settings of a website and server, you might be able to adjust this setting as well.  Once we did so, we were able to host images of the selected texts on our institutional instance of Omeka, which then interfaced very nicely with Scalar. Indeed, Omeka is one of the "other archives" options pre-populated into the Scalar interface, in addition to their affiliated archives.  

With our digitizations available to embed, we then turned our minds to analysis and curation. Each recovered excerpt was the centerpiece of its own "chapter," contextualized by research-based critical commentary and other relevant media located through online searches. For example, the chapter on Mary Wollstonecraft features a rare book seller’s description as embedded media. Groups varied in their approaches to layout for the embedded selections, where annotations may or may not appear, and the embedded media for some selections would be at the start of the page, or after a short summary.

Within their respective “chapters,” student-authors primarily represented their selections and other component media through Pages, which were the primary anchors for content sharing. These pages were not uniformly named or organized, but instead reflected the priorities and organizational schema of individual authors and project groups, representing the selected recovered texts in widely varying ways. In fact, there is a considerable difference between the lengths of selections, the file type used, and other representational choices. Pages dedicated to Eliza Bradley, for example, featured multiple excerpts from Bradley’s larger work, instead of multiple pages constituting a single excerpt, while multiple excerpts from Frances Cobbe totaled around 60 pages embedded as a PDF frame, allowing a user to read the excerpt through an Adobe viewer without changing pages on Scalar. 

Several groups utilized annotation features to provide context and interpretation of their artifacts. Annotations allow close textual work that attends to specific discursive features of a given text or media object. While we later learned about the hypothesis plugin available for Scalar annotations, we did not have experience with it in our work, and our resulting projects use other varied approaches that are probably less effective.

One thing we discovered through experimentation with various file types and annotation features is that PDF files, which allow for Optical Character Recognition (OCR), are not compatible with Scalar’s annotation features — a lesson we learned the hard way in our own trial-and-error process. Instead, the annotation feature groups text with the use of a click-and-drag box dragged over any area of the file, treating the files more traditionally as visual images. The inability to shorten the pop-up windows with annotations could cause reader confusion, such as in Winnemucca’s excerpt. As a result, a reader might be confused by large annotation boxes that refer to a general area or overlap with a note that appears and would be likely to be even more confused if the annotation is designated as a path itself, which is an option within Scalar. 

As a result of different file presentation choices, the scope of each analysis widely varied. Some pages discuss the text holistically, like through the analysis of writings both by and about Phillis Wheatley, for example outside a contextual “close read.” Others annotate the selections with commentary highlighting larger themes and connections. Still, others transcribe the text from outside the image and onto the webpage (as in the case of the Women’s Recipe Book, which was in manuscript form). 

We used other content types, such as visualizations, at the project level. To achieve cohesion, we also collectively decided to have an introductory splash page before each, which would appear as a "part" in the Table of Contents. These gestures were important to use to demonstrate an overall unity to the an(ti)thology as a whole.