Dr. Scannell’s project
The project involves analysis of socialgraphs, in particular in the context of politically-divided or"polarized" communities. I collected a dataset of twitter users fromBelfast (and who they follow) over the summer that a student would beable to look at as an example. I can describe things in more detail to any students that might be interested.
Dr. Flavio Esposito’s projects
Project P1: Design and Implementation of a FYREFLY Architecture Component
Fog Computing is a paradigm in which much of the processing takes place in a process at the edge of the network (as opposed to in the core of the network as in the Cloud Computing paradigm). This distributed approach is growing in popularity because of the Internet of Things (IoT) and the immense amount of data that sensors generate in many industries. It is inefficient to transmit all the data a bundle of sensors creates to the cloud for processing and analysis; doing so requires a great deal of bandwidth and all the back-and-forth communication between the sensors and the cloud can negatively impact performance. Typical examples of Fog Computing applications are urban mobility-on-demand systems, networks of drones for surveillance and environmental monitoring and disaster scenarios, fleet of robots for agriculture or healthcare applications, as well as autonomous aerial vehicle systems that provide connectivity to ground stations in rural or disaster areas of our planet. A common challenge across these networked applications is the ability to provide a persistent service responding to events that may change the state of the network, e.g., node or link failures.
Based on skills and interest, in this project the student will be required to participate in the design, modeling, simulation or even real implementation of a management application (software component) for FYREFLY, a novel Fog Computing architecture, using, e.g. ROS 2, a Robotic Operating System designed for distributed robotic systems or ONOS, the Open Network Operating System.
Project P2: Randomized Consensus for Reliable Software-Defined Networks
In a Software-Defined Network (SDN) scenario, a controller is responsible to read, modify and adapt virtual switches across an entire network. One problem that SDN system designers currently have is how to reach asynchronous consensus across multiple controllers to manage the (multi-provider) network cooperatively and in presence of arbitrary (realistic) failures. Arbitrary failures, also known as Byzantine. It is well known that it is impossible to solve the consensus problem deterministically in an asynchronous system if even a single process fails, since in such systems is impossible to distinguish a failed process from a slow one. The student is required to design, implement and evaluate a randomized distributed consensus protocol for the Multiple Controllers Byzantine Consensus Problem, leveraging the code available within the Open Network Operating System (ONOS) community.
Project P3: Migration of a Fog of Docker Containers
Docker is a technology that is gaining tremendous popularity and adoption among tech companies. One of the challenges of Fog Computing is to efficiently migrate multiple processes to always provide a reliable service even when the underlying physical network is unstable, for example in cases of (natural) disasters, or in case of proactively preventing attacks. Instances of Docker containers within the hosting Fog Computing infrastructure may boost the elasticity property of such Cloud-based services, given their speed of deployment. In this project, students are asked to perform a performance analysis (with a measurement campaign) of existing (live) migration solutions for a network of Docker containers i.e., a Docker Swarm. Such measurement campaign should motivate the need for a better migration strategy. Students may also implement a live migration strategy in real Linux settings, and release a downloadable testbed using fast prototyping tools such as Mininet.
Project P4: Optimizing Live Migration of Fog Networks
One of the challenges of Fog Computing is to efficiently migrate multiple processes to always provide a reliable service even when the underlying physical network is unstable, e.g., due to a disaster scenario. Instances of virtual machines or Docker containers within the hosting Fog Computing infrastructure may boost the elasticity property of such Cloud-based services, given their rapid deployment and scalable properties. In this project, students are required to devise and solve a novel optimization problem whose goal is to optimize the bit rate allocation for the live-migration of multiple Fog Computing virtual instances running on Linux containers, minimizing the total migration time. The optimization problem will then be solved with classical small scale solvers, e.g. within MATLAB. Heuristics should be then proposed and implemented within a real Docker environment.
Center For Digital Humanities project opportunities
The Center for Digital Humanities at Saint Louis University has a wide variety of projects
suitable for capstone projects. Our staff has experience as client, subcontractor, and mentor in a
variety of projects and will provide the support required for success.
About the CDH
The technological focus of Digital Humanities at Saint Louis University has been to develop
foundational, reusable, and interoperable technologies. The open nature of OAC and IIIF allows
for relatively easy migration of data in the long term. Projects built by the center will not be made
redundant by new technologies, and moving to new technologies in the future is not restrained
by proprietary or legacy systems thereby meeting need for sustainability. Most code is hosted in
open repositories and includes contributions to open community projects, ensuring that student
work will have a life outside of Saint Louis University and may become legacy code in important
projects.
The CDH is focused on building applications to support humanities research and pedagogy. We
have built applications for transcription, critical edition creation, exhibition of multimedia
scholarship, learning paleography, and reconstructing dispersed materials virtually into their
original configuration. Our work has been done in support of emerging standards such as IIIF
with collaboration and consultation from international universities, museums, and libraries. We
maintain an open public IIIF store for smaller institutions to be able to get going in Humanities
projects.
The Technology We Use
● Web application and user interaction (HTML5, AngularJS, jQuery, CSS3)
● Servers and databases (Java, MySQL, MongoDB)
● Open Standards for data modelling and exchange (IIIF, SharedCanvas, Open
Annotation Collaboration, JSONLD)
Project Opportunities in the CDH
These may be bundled to fulfill the requirements of a capstone. Each of them is scalable or
decomposable to the skill level of the assigned team or student.
1. TPEN (Transcription for Paleographical and Editorial Notation)
TPEN is our most successful application. We are approaching five years of operation and there
are number substantive features that we would like to add to the application as we work towards
our 3.0 release.
A. Alternate interfaces for variable content
TPEN transcription was initially designed for rectilinear left to right text in traditional codex
proportions. This capstone project would involve building out singlepurpose
interfaces required to handle images of other materials within the TPEN ecosystem. Any number of targets may beset for this project. TPEN is already equipped with endpoints and modules for users' andresources' CRUD operations; this project focuses on proliferating singlepurpose
transcriptioninterfaces by creating the UI/UX with some design collaboration with the CDH to incorporate anynovel data structures into existing standards.
The target resources include variations such as:
● Format—scrolls, newspapers, folio, maps
● Textflow—right to left text, left to right text, vertical text, spiral text
● Content—nontext
resources, such as music or landscape features
● Marginalia—commentaries, glosses, hand written notes and other text on their pages
beyond the main text.
B. TPEN for Touch interface
The vanilla TPEN interface is reliant on the desktop experience, as it was developed before any
touch device had the power or resolution to complete even the simple tasks required. However,
a clever designer may not only find opportunities in enhancing the generic interface to a
touchenabledone, but may also envision smaller screen experiences for TPEN that will allow
project management, sharing, or viewing with reduced functionality.
C. Content Access and Delivery
TPEN exports in plain text and xml fragments. It would be advantageous for the application offer
export as properly formed TEIXML documents, IIIFJSON,and other formats. This project
would require knowledge of Java and interaction with our existing backend code base. The
CDH will provide support for standards discovery and navigating any problematic code.
Also within this category is the integration of APIs for remote image locations. Currently, TPEN
allows users to upload a zip file of private images to be hosted for transcription. The CDH
intends to discontinue this hosting service with preference for API connection to Internet
locations such as Dropbox or Flickr and the use of local storage to access.variety of sources, such
asdropbox.
D. Tagging upgraded to Brat style annotation
TPEN’s current transcription interface does not allow for a rich user experience when annotating
as the BRAT tool does found here: . This would be a dropinreplacement
for the textareas in TPEN, but could be developed as a standalone tool in the first instance. It is
expected that the design will require facility with SVG elements. This would greatly improve the UX for annotation in TPEN. All the connections to theannotation store APIs need to be plugged into the tool, but are available already.
E. Manuscript Image Analysis
Improve image auto parsing, edge detection and image rotation. Currently TPEN does some
rudimentary image manipulation to autodetectcolumns of text in the images to aid in
transcription. A more fully developed image parsing tool would be of great value to users.
Build an interactive application module for reuse in the Open Annotation Collaboration and
International Image Interoperability Framework communities for region identification on image
resources. The primary use case for this application is within the TPEN application. Successful
projects will be able to identify the page, text block area, column separation, and lines of text on
digitized images of manuscript resources of various sizes, colors, and resolutions. TPEN
currently implements a simple parsing engine written in Java, which may be extended or
replaced. This method may be implemented as a web service with the identified regions
returned to the requestor as a SharedCanvas compliant JSON object or as a clientside
algorithm. Further development options, depending on the requirements of the capstone, should
include user interaction to adjust the result, including at least a minimal HTML application
interface. This project has high visibility within DH and Medieval Studies and cuts across several classesof skills. The CDH has an available datastore for primary documents and annotations, so theproject focuses on the image analysis elements.
F. Interface Customizations
a. Design compatibility for language packs and relevant text input direction.
b. Implement translation and transliteration layers on transcription annotations.
c. Accommodate alternate views of transcription along with querying personal and public
transcriptions.
G. Mirador viewer into TPEN
TPEN will hand its project manifests to Mirador, a SharedCanvas viewer, and Mirador can
render all the canvases and annotations to the user in a well tested and maintained
environment. A successful inclusion would be a standalone page where a user could view their
project in the viewer (a project includes a JSON sc:Manifest).
2. RERUM.io
RERUM began as our open annotation store and now hosts a selection of IIIF compliant tools
that can be used together or separately for various scholarly purposes. There are a number of
existing applications we have or modules we wish to enable on rerum.io:
A. Textual analysis tools
The annotation store contains textual data from many sources and the tools would query and
analyze the data to run useful textual analysis tools. Many of these tools already exist in open
source communities, but are largely commandlinedriven and would need to be integrated into
a useful web interface.
B. Text annotation interface
The annotation store stores sc:Canvas objects, and as such we would like there to be an
interface where a canvas could be loaded to be annotated directly. This could also support
users giving the URL of a canvas from a different server and attempting to annotate it as well.
We have an interface already started to build from in our application Tradamus, but an entirely
new interface could be developed as well. This project would require some collaborative design
because the standards for text annotation are not clearly established yet. It is also likely this will
require interaction with several other scholars (all English speaking) who are addressing similar
challenges.
C. CRUD operations
Implement CRUD operations UI for users to manipulate their objects directly from the front end
connected to the annotation store. The more comfortable the student is with design, the more
complex the possibilities become. The basic API for these actions are already established, so
no backend
coding will be required.
3. Fragmentarium
There are a large number of manuscripts and documents that have been separated, ripped or
deteriorated over the centuries and Fragmentarium seeks to allow users to bring scanned
images of these fragments together in a straightforward interface to reconstruct the pages of the
manuscripts. This would mean bringing many images together onto one canvas and putting that
canvas back together like a puzzle. The pieces and the canvas must be rotatable, editable,
clippable, scalable and moveable. The resulting object must be represented as a valid IIIF
object.
4. RTI Viewer
RTI is an imaging technology developed by HP which allows for dynamic relighting of lowrelief
objects. There is currently a Java program for viewing the resulting images in a desktop
environment, but we seek to create a web viewer for RTI images. This requires openness and
code sharing from the RTI Viewer publisher, but building a web viewer from the Java code
should be straightforward and is important to allow pseudo 3D viewing of objects. In a complete
project, the relatively large image files will also be redesigned to accommodate web delivery at
lower resolutions.
5. HISGEN Genealogies
The CDH has a very large database of Irish nobility for the preChristian
period to the late
medieval. This collection needs to be digitized and to facilitate both creation of the database and
leveraging the relationships inherent in the material the Eventitiesdata model was developed. In
support of this goal, a creation interface for new events and entities needs to be created and
tied to existing image annotation tools to enable the scholarly encoding of the images in this
collection.
6. Heritage Hero Mobile App
As has been demonstrated by the ongoing war in Syria and the extremists in Mali important
cultural heritage sites and archives can be threatened in ways that UNESCO and others cannot
react to quickly enough to help. In Mali, French military intervention has allowed the Mail
government to work with the Hill Museum and Manuscript Library (HMML) to do a large scale
digitization project to preserve the unique collection of manuscripts, documents and objects in
Mali. Such an effort takes time and the equipment is is often not available, as demonstrated in
Afghanistan and the Middle East most recently. Existing websites and applications index official
sites as travel destinations, but are not open to contribution or scholarly annotation.
This app would allow people to photograph and record metadata about any historical object or
location and upload that data to our annotation repository. The app will exchange data with the
repository as IIIFcompliantJSON objects (oa:Annotation). The best project will also include a
desktop application for managing and exploring this data. This project connects with initiatives
at University of Pennsylvania and represents an opportunity for the capstone developers to
have input on the early planning of the project, if desired.