Library DigiHub – Providing Accessing to Institutionally Hosted Virtual Computational Services and Applications

Over the years, I’ve tried to imagine various ways in the Library might engage with what I’d call digital skills rather than what they claim as such, so I was completely taken aback today to learn about the new DigiHub. (I guess this in named in part on homage to DigiLab. Although now just a meeting room, the original vision for the Digilab, at least as I remember it, was also that it was a space where you could try out new and emerging technologies and devices; it included several games consoles, for example, as well as some of our tabletop programmable Lego robots).

From what I can tell, the DigiHub service will comprise of three main parts:

  • a digital application shelf (DASh), that will allow patrons to run Docker containerised applications on a Library server or on their own computer. The applications will include both course related applications as well as “generic” applications such as RStudio or OpenRefine; for unauthenticated users, this will launch temporary servers; authenticated users will also be able to save and load files from their connected institutional OneDrive account; the DASh will also include a historical items area including runnable versions of software from courses pasts. (I wonder if this is based on the Emulation as a Service  that I gave a quick review of here three years or so ago now? I also notice that several other institutions already host related things, such as CMU’s Olive Executable Archive. See also the Software Preservation Network?)
  • a digital integration workbench service (DiWb (pronounced “dewb”?!)) that will use Jupiterhub to provide access to integrated services from a single login. From what I can tell, these will be prebuilt environments such as that exposed by the TM351 VM that interconnect multiple services (eg a Jupiter notebook server and a Postgresh database server) within the same environment);
  • a “digital replay” service  (DRS), linked to items in the ORO research repository, that will allow patrons to “replay” notebooks or R scripts associated with repository items. Again, this looks like a ‘small pieces loosely joined’ approach that has been put together from already existing pieces: a Githob repository (or maybe a locally hosted Gitlob repository?) is associated with publications that include reproducible research code items and project data files, along with a declaration of the software packages, and version numbers, required to run the software. Binderhob is then used to launch and run a temporary environment where code and analyses associated with a particular paper can be run and further explored.

This is all really exciting, and beyond my wildest dreams of what I thought the Library would be able to achieve within such a short time of getting into the idea providing digital computational services. (I can’t wait to see how they develop subject specific areas, for example, hooking in to the  OpenSTEM Labs, perhaps creating a Digital Humanities Lab, and maybe even helping my department see why a Digital Computing Lab would be a Good Thing…) In the meantime, I guess I’ll be able to play with DASh and DIWbs. I also wonder how many papers currently in ORO are in a position where they can be augmented with replayable research scripts, or whether this was a pre-emptive step from the Library seeing how researchers elsewhere are starting to support, and promote, their publications with associated research assets?

So does this represent the latest step in the evolution of technology use at the OU? (Related: Martin Weller on technology in From the University of the Air to the university everywhere.)

According to project manager Joanne Kerr, DigiHub services will start to appear later this month; the first will be a temporary notebook server, followed by the soft launch of the Digital Application shelf with items relating to courses that already run containerised, or easily containerised, software applications and a demonstration of some notable applications from the OU’s past on the Digital Software Archive Shelf. The repository runner will be out later in the year as part of a Digital Hub (RIS) (research infrastructure and support) foollow on project.

Author: Tony Hirst

I'm a Senior Lecturer at The Open University, with an interest in #opendata policy and practice, as well as general web tinkering...