Regions of Interest Quantified – Solving Problems of the Analogue Methods of the Discoveries in the Judean Desert Series (Part III.3)

Before we return to solve some of the issues of 1QSa II 11–12, I think it would be helpful to reflect on the importance of the previous editions of Discoveries in the Judaean Desert (hereafter, DJD) with respect to the best practices introduced in the previous post. Specifically, I would like to examine how digital editorial methods improve on what I am calling analogue methods.

In the previous post of this series,1 I stated and described the importance of implementing the ideas of best practices. As we move away from the analogue editorial methods of the DJD and move into a digital framework, there is a need to enhance and improve upon the editorial methods of DJD. These enhancements and improvements further contribute and add to the value of the DJD editions.

I introduced two best practices in the last post:

  1. Collect all images of any given fragment
  2. Maintain the integrity of the Original Image

One could argue—and I would agree—that both editorial practices were utilised in the production of the DJD series. To be sure, we could skim through any edition and see an editor appealing to different images to make their readings—this is not new. But collecting every image and tagging it in a database model provides a robust way to create Palestine Archaeological Museum Stacks (hereafter, PAM Stacks). A PAM Stack is a

complete historical inventory for each artefact, as it was imaged in the sequence of the early PAM images and the newer IAA images. This entails that each fragment on every PAM receives a unique identifier, as well as the PAM plate itself. Once the PAM plates have been tagged, we bind the PAM identifiers to the singular PAM image used in the editio princeps and furthermore bind these to the colour image of the IAA, which is designated as the master image.2

My improvement to this method is to account for the images in a relational database—and this is new. By engineering a database model, it is possible to track the images, thereby making it very easy to access each and every image, and display the PAM Stack in an HTML5 Canvas. Here is an example of the database model, I have created:

Database Relations of PAM Plates

The above diagram represents five tables of my database.3 The far-left table, pam_plates, is a full accounting of the PAM images; the table in mid-center, pam_plates_frags, is the Region of Interest (ROI) of every fragment on a PAM Plate;4 This table holds the coordinates of the fragments on the PAM image, so that it is possible to crop the fragment from the PAM with algorithms. In other words, the integrity of the PAM image (as a plate) is maintained (best practice no. 2; see above); The lower-center table, pam_frags_human, are the editor’s siglum assigned to a fragment, which was published apart from DJD. There are only a handful of these. In the last post, actually, we saw an example of this. The next table on the lower-right, rois_to_pam, are Regions of Interest on the fragment itself. These Regions of Interest (ROI) can be used to annotate the fragment (i.e., capta), whether it be traces of ink, characters, worm holes, stitching marks, line rulings, material damage, etc. The upper-left table, iaa_to_frags, links the Leon Levy Dead Sea Scrolls Electronic Library image to the PAM Images.

The above model creates many benefits that extend well-beyond the DJD editorial structure. First, the above model reduces the amount of time required to search for a various image throughout the PAM images and to record what image was influential to an editorial decision.5 Based on the above data model, I have been creating “fragment editions,” by which I mean an edition that focuses on the fragment itself. An example can be found at my Github page. These fragment editions are only part of a larger digital edition, that concludes—when possible—with a 3D edition of the Scroll. The reason why this series is called On 2D Material Reconstruction of the Dead Sea Scrolls is because the best practices I articulate here for 2D reconstructions are preliminary steps to make a 3D reconstruction!

Second, the above model creates an easy way to maintain the integrity of the original image. Why does it matter? It matters because the digital medium affords sophisticated measures to analyse the fragments, so as to gain a better understanding about palaeography, historical linguistics, and material reconstructions. It is necessary to maintain the ability to regress to the original image. We will see this in practice in the next post.

In the past several posts, we have been focusing primarily on palaeography. In the next post, we will finally return to 1QSa II 11–12, by using all the PAM images, to see if we can get a better reading of the these two lines.

  1. All of the posts in this series are listed on the Second Temple Judaism resource page.
  2. James M. Tucker and Peter Porzig, “Between Artefacts, Fragments, and Texts: An Analysis of 4Q266 Column I,” DSD 25.3 (2018): 341.
  3. I engineered this database to make my editions for my doctoral thesis, From Ink Traces to Ideology: The Material, Text, and Compositional Development of Serekh ha-Yaḥad with New Critical Editions.
  4. See this post for how I tagged and parsed the PAM plates.
  5. I gave a presentation on these ideas in Göttingen in 2017. See here for further details.

Leave a Reply

Your email address will not be published.