Doing archival research in the 21st century has many advantages. We are no longer confined to working long hours in windowless viewing rooms with small golf pencils and paper to take copious amounts of rushed notes on primary source documents, historical objects, or works of art. Instead, we can go in with our smart phones, take photos of what we need and digitally “bring” these sources home with us for thorough inspection in the comfort of our offices or kitchen tables. Moreover, many of these archives are now accessible online, in the form of high-resolution images, 3D models, and XML (Extensible Markup Language) documents, bypassing the need to travel long distances or wait for rare volumes or objects to be made available. These conveniences, however, often come with a caveat: the sheer volume of sources and images that we collect in the research stage of our projects has made it almost impossible to keep track of what we’ve seen without a proper organizational system. Left to accumulate organically, this “archival bulk” can significantly slow down the research and writing process and inevitably leads to hours of frustrated searching and lost work.
During my dissertation field work, I amassed an unsavory number (upwards of 10,000) of images documenting sites and objects. Finding an intuitive and customizable personal image management software suitable for the uses and needs of a young scholar (i.e., free or low cost) was imperative and surprisingly challenging. This issue is especially pressing for art historians and other academics who rely on or study images and objects because we often have personal databases with several tens of thousands of files. When I asked colleagues and a digital services librarian for advice back in 2017, I received a lot of plaintive echoes of confusion and desperation and a few provisional recommendations which I will include at the end of this post.
Last year, with more time on my hands, I revisited this issue and discovered Tropy. Developed at The Roy Rosenzweig Center for History and New Media at George Mason University, Tropy (like its sister program Zotero) is free, open-source, and fairly easy to use. I will preface by saying that Tropy is not perfect, but it is a customizable, light (less than 100mb), and stable desktop software that I have found to be extremely helpful for organizing the images and sources that populate my dissertation.
Tropy arranges collections into “projects.” The screen capture above shows my Tropy project of a new text that I’m working on now. The panel on the left shows you the project title, the folders within the project which Tropy calls “lists”, and tags that I’ve created to help me catalogue my images and keep track of my research progress. The middle pane contains thumbnails of all the images in a selected list (in this case “Tonghwa-sa”), and on the right are three information panes for metadata information, a larger, complete thumbnail of the selected image, and at the bottom a list of specific notes that have been attached to that image. A great feature of Tropy is that it does not duplicate your image files but does not affect the originals in anyway if you edit them within Tropy. So, rest assured, anything you do in Tropy will not affect your original images.
One of Tropy’s main selling points is its flexibility and low barrier to entry. When implementing a new cataloguing system, often the most tedious step is meticulously pre-planning the structure of your data. But with Tropy it is entirely possible to jump right in and adjust the organization of your project as your thinking develops. Lists can be renamed, nested, moved, and adjusted by dragging and dropping. Metadata templates that help you identify your images can be customized to better fit your research questions and objects. Tropy comes with three templates that are designed for documenting archival texts, because that is what it was primarily built for, which included fields that were not relevant for me as an art historian. However, I was able to easily add and delete fields to create custom data that are useful to me. For example, a field that I added for my own GIS-dependent project is elevation range which I can query when I want to identify all images that are currently located at certain elevation above sea level.
Customized metadata template (left) and a populated metadata pane (right).
Another significant feature is the ability to group several images of the same object together so that the contents of my lists are more obvious in one glance. Metadata can also be batch edited which saves a lot of time when you have 50 images of the same object from various angles. Double clicking on an image will take you to an object window where you can add observations, references and citations, and other notes specific to an image. Here you can also highlight a section of an image and take further notes just on that aspect. These notes are all searchable and quickly accessible from the main screen.
Selection tool and annotation in Tropy
Finally, the tags can be used to identify image groups that are not part of your original list structure, such as by theme or research avenue. Tags add an extra flexible layer of categorization that you can use not only for content keys but also for workflow. For example, I have a “needs transcription” tag that lets me know which objects or images need to be transcribed before I start writing.
Since its launch, the developers of Tropy have updated the software to build out its functionality and incorporate user feedback. Significantly, Tropy now supports a larger array of image file types (older versions did not support PDFs) and you can upload Tropy projects (as .tpy files) to a cloud-based folder to work collaboratively. However, Tropy does have a few serious drawbacks. It does not support mp4 or video files, so if your sources involve multimedia or audio, this may not be the tool for you. For images of texts that have an OCR layer, Tropy can’t yet access that layer and it does not have built-in OCR. Its export functionality also leaves a lot to be desired. Currently you can only export your metadata and notes to JSON-LD, or to a CSV or Omeka S after installing the requisite plugins. Tropy does not capture metadata from images downloaded from an online archive and it does not create citations from its metadata. A last weak point is also one that Tropy developers are working on: the ability to link citations and data on Zotero to images in Tropy. Having cross-platform connectivity would solve the citation and capturing metadata issues and allow scholars to really integrate these tools into their workflow.
Despite these limitations, Tropy is a robust tool that allows scholars to productively organize and annotate images for scholarly research. It has great documentation and a responsive support forum. For humanists who are interested in taking advantage of the potential of incorporating metadata to their image archives or want to streamline their image-based workflow, Tropy is a worthwhile investment.
Some other alternatives for those who are not sold on Tropy:
ArtStor: For those at institutions that subscribe to ArtStor, one option is to set up a personal image collection where you can add images and describe them within a very basic metadata template and search them using ArtStor‘s big search engine. This method however is imperfect because it requires internet connection, and the size of personal image collections are capped.
Zettelkasten: Another software that was suggested to me was Zettelkasten – a free, no-frills German program based on the principle of the personal index card system used most famously by Niklas Luhmann. Though the simplicity of the system was attractive, it is primarily built for text and so is not a purpose-built solution, but one that may work for scholars working mostly with text.
FileMaker: Accepts images and allows you to develop your own metadata template. It has a pretty loyal following. However, it is prohibitively expensive and not particularly intuitive. There is a 45 day free trial, so you could give it a test run to see if it’s worth the money.
ARIES (ARt Image Exploration Space): ARIES is a digital initiative begun by members of the Digital Art History Lab at the Frick Collection in New York. The online tool builds upon the analog capabilities of lightboxes as used by art historians when conducting comparative analysis. This system was clearly developed for and by art historians because it offers very simple methods of creating the types of visualizations that art historians often rely on such as timelines, grouping of images, comparing small sections of an image, comparing relative sizes of artworks, overlaying semi-transparent images on top of one another, etc. This one is great for those looking to easily create digital versions of such visual analysis exercises for online classes.