Armenia

Digital Humanities in the Field

Chris and Colin hard at workChris and Colin hard at work

by Christopher Besser, Master’s degree graduate, University of Pennsylvania
and Colin Roberts, Bachelor’s degree graduate, University of Pennsylvania
June 27, 2018


Barev! Chris and Colin here! We are both software engineering graduates working on the 2018 Open Archaeology project in Armenia, as the chief developers of several of the digital technologies in use in the field. This project has been a lot of fun! In many software projects, the developers only interact with the end users through intermediaries, be they online feedback, requirements engineers, or email. However, with this project we the developers are field testing our own apps alongside the end users. We are personally able to determine what works, what doesn’t work, what needs tweaking, and what additional features we should add. We feel this direct interaction has both facilitated faster implementations of fixes and features, as well as allowed us to grow as software developers by improving our ability to anticipate and address problems the users may face before they actually encounter them.

Data Collector at workAndrew using the Data Collector app

Leading into and during this summer, we have worked on three key software projects for use in the field in Armenia, two Android applications and a web service.  Our work prior to coming to Armenia has been supported by the Price Lab for Digital Humanities, and our trip to Armenia was made possible through student travel grants from the Penn Museum. The first application, Location Collector, communicates over WiFi with an Emlid Reach, a highly accurate differential GNSS device, for extremely precise location data on objects found in the field. The app can also revert to a phone's internal GPS for people without access to a Reach. This app allows the field-walkers to photograph objects in situ. Whenever the app has an internet connection, the user can upload finds data to a central database. The second app, Data Collector, is for use in the lab. This app communicates over Bluetooth with a digital scale, as well as over WiFi Direct with a Sony-API-enabled camera. Users can look up objects previously registered by the Location Collector app, record their weights with the scale, and photograph them with the remote camera. The app then downloads the photographs from the remote camera and neatly organizes them on the Android device’s file system for later upload to a cloud server. For users without access to a Sony-API-enabled camera, the Android device's camera may also be used. Because these applications need to write data about objects to a remote database, we wrote a web service in Python Django, the most popular Python-based framework for web development. This web service handles object uploads from the Location Collector app as well as data uploads and downloads from the Data Collector app, so neither app needs to directly query the database.

Colin enjoying natureColin taking a break from coding to enjoy nature

As is the case for most interesting software projects, we encountered and resolved several challenges in making everything work. The Data Collector, for instance, is intended to be easily adaptable to any Sony-API-enabled camera, not just the Sony QX1 being used in field. However, according to Sony’s API reference guide, there are nine different versions of the API, and each particular camera implements only one of them. Additionally, certain cameras have slightly different usage requirements for the same API calls. Meanwhile, besides the QX1, we only had one other camera to test on, a Sony Alpha 7. To be as sure as possible that our app will easily adapt to other cameras, we had to carefully consult Sony’s 286 page API manual to make sure we covered every possible scenario.

An additional complication we encountered with taking photos with the Sony API is that the camera captures large images of the size 8MB or larger.  This requires significant resources from the Android device to download and save the image files to local storage. Generally, Android applications load thumbnail images a couple hundred KB in size, so displaying images in this manner is not normally problematic. We wanted to make sure we were working with the full images rather than thumbnails, so we had to carefully manage Android resources. While Java usually handles memory allocation and recycling automatically through the garbage collector, we needed to manually allocate and free memory like we were writing in a lower level language like C or C++.

Challenges aside, we developers were incredibly lucky to be working on such a cool project. Nothing beats looking out at the beautiful mountains of Armenia or eating fresh-off-the-tree apricots as we work. I’m incredibly thankful to Price Lab, the Penn Museum, and all of its donors for making this trip possible, as well as the University of Pennsylvania, Dr. Cobb, and The Armenian Institute of Archaeology for their help with lodging, travel, and other assistance. Last but not least, we would like to thank our friend and translator Hayk jan for his help in the field and for helping us make the most of our Armenian journey.

Chris looking out at the Armenian mountainsChris reflecting on his time at Armenia



Links to the project code repositories:

This page has paths:

  1. 2018 Fieldwork in Armenia Peter Cobb