This will be a roughly tri-weekly update, which equates to my time allocation to C.E.
I have spent a bit of time putting final touches to my long blogpost on the Trade Directories work, and then trying to promote this a little bit. It seems to have caught a few people’s interest online, including an eminent historian who’s keen to work digitally and I’ve had a few contacts through this already.
I’m involved in the ongoing task of pulling together the documentation produced during the December workshop into several forms:
A second and much more detailed blogpost that will outline the computational approaches available to pursue this challenge further. The idea here is that this can then be something people, we, I, can point to when wanting to explain the ambition of this investigation.
I’m currently editing together contributions from various Turing colleagues into a text which will go up on the Living with Machines blog in due course — and may even need to be split into more than one post, since they prefer shorter nuggets. This will be nice once it’s done as we can get the CE blogpost to point there, and that post will link back to the CE blog, which will give some coherence to the shared work.
I’m also involved now in drafting a proposal for a larger package of work using multimodal CV to parse and analyse ‘difficult’ historical documents. The idea here is that the Leicester TD collection offers a perfect testbed for this kind of technique, because it’s a singular collection — of serious thematic interest to C.E. — but also contains enough visual heterogeneity within it to be sufficiently challenging as a data science problem. It’s a fine balance between the specific and the universal.
In the first instance, this proposal will be targeted to Turing’s Research Engineering Group, in the hope of persuading them to accept this challenge, which could then become a collaborative investigation within C.E., with technical leadership from the Turing. I’m conscious that there are several possible barriers to this happening, and so there might be ways to adjust the scope of the investigation so that it fits better within the envelope of what C.E. might realistically take on and support; as it were, as a back-up plan. I expect to bring a draft of this to a digital investigations meeting before long, but am very happy to have input from colleagues immediately as I start drafting this over the coming weeks.
BT archives have now agreed to provide us with an export of their database metadata and I’m in ongoing discussions with James Elder about the format and nature of this material, which I’m looking forward to receiving shortly.