As we continue work on the redesign of the Archive, our collaborative efforts with programmers and web designers who are unfamiliar with Blake’s work reveal aspects of the Archive’s structure and organization that we take for granted. Our bi-weekly meetings often involve volleys of patient explication: the Blake folks (Joe Viscomi, Ashley Reed, Mike Fox, and myself) offer mini lessons on Blake’s multimedia production in order that the designers and programmers better understand the content they’re working with; and they in turn lecture us on the possibilities and constraints involved with the database structures, modified programming languages, etc. that will display that content.

One recent sticking point has been the concept of virtual groups. Virtual groups designate objects that Blake himself never conceived of, produced, or bound as a unified work in the way that he grouped, say, the eighteen plates of America as a single work. Examples of virtual groups include the Large Color Printed Drawings, the Pen and Ink Drawings, the water color and painted Illustrations to the Bible, and Blake’s letters. These groups are titled—and their constituent objects chosen—not by Blake, but by the editors of the Archive, and thus they are a blatant example of the editorial intervention that necessarily happens in a project like this. There are currently nine virtual groups in the Archive, but they are nowhere identified as such to users. We have tried to display virtual groups in the same way that we present other works, but as we develop the redesigned site, the ontological differences between virtual groups and the rest of the works in the Archive have made treating them in the same fashion impossible.

The fundamental difference lies in the fact that the “objects” in a virtual group are separately titled (by Blake) works, like Nebuchadnezzar or The Number of the Beast is 666. This is an exception to the prevailing hierarchical architecture of the Archive, originally designed with illuminated books in mind, which moves from the abstract work (The Marriage of Heaven and Hell), to the specific copy (copy H), down to the particular object (object 11)—the latter of which is featured on the OVP (Object View Page), the fundamental unit of the Archive. In terms of our data structure, all the editorial information for each multi-object copy (as well as the order of the objects) of works like the Marriage is contained in one file, the BAD (Blake Archive Document). Thus, the information that the OVP displays for object 11 of the Marriage copy H can be neatly drawn from one section of one file. But for virtual groups, since each constituent object is in fact a different work (with a different title, catalogue number, date of printing or composition, etc.), each requires its own BAD file. These files are “virtually” aggregated, which in turn necessitates extra files (designating the order of virtual objects) and programming to find and display the pertinent information.

Since the pertinent information for virtual group display and functionality is scattered across a range of documents, as we redesign the site we have had to develop and share spreadsheets listing the crucial display information for each work in order to aid the programmers. And in addition to group emails and our regular meetings, we have had to sit down with members of the programming team and work out various kinks associated with the virtual group display. These interactions have also revealed to us some (minor) inconsistencies in our own BAD files. Sometimes, for instance, we include the “The” of a title separated by a comma at the end of a title:

Screen Shot 2016-03-23 at 1.04.41 PM

In other places we don’t:

Screen Shot 2016-03-23 at 1.05.18 PM

No wonder the programmers are confused when we say that the title of The Book of Thel shouldn’t display as “Book of Thel, The.” (In our defense, when you have as many people involved in a project that has been going on for as long as the Archive has, one is bound to discover inconsistencies.)

Working out the details of virtual groups in the redesign process testifies to the oft-repeated claim that digital humanities projects are predicated on collaboration between people with different areas of expertise. And this collaboration often works best when people stop emailing each other, or talking in meetings, or filling out tickets on GitHub, and decide instead to sit side by side in front of a computer in a windowless office and start touching digital things with their human fingers.