My updates and work-in-progress

This will be a roughly tri-weekly update, which equates to my time allocation to C.E.

• 1 Mar 2023
  • I received a large export of BT archives open catalogue data: sets which include both the imaged material, as well as the whole collection, which I wanted for the sake of comparison. This is in XML format. Although I requested a CSV as well, James advised that TSV had worked better in the past from their Axielle system. Sadly this file isn't really readable so I will need to wrangle the XML file into something more legible when I get the chance to get it into a Python notebook, which is on my to-do list.
  • I'm keen to locate and gather any extra material or documentation about this database and hope to get the chance to speak to James about this once Jon Agar and I have had a discussion, currently planned for tomorrow. I'm reluctant to dive into using this material until we can establish a little more about some of the selection criteria; to which end, I may try to contact the previous archivist who was involved at the time of the digitisation. Once I have a better sense of how possible all of this might be I can upload the original files into Box if others want to have a look at them.
  • Jon and I have been in touch with a PhD student in architectural history who has been working on BT exchanges, who has a connection to Historic England, so we're aiming to have a chat with her later in the month when she's available to see if there's a productive overlap with material we may have access to.
  • In more general project matters I'm excited to do some reading around the notion of a 'social machine' which I understand is the basis of the meeting planned for Leeds. I'm hopeful I'll be able to make at least one of the two days, although sadly I'm already committed to a workshop on the Tuesday in London.
  • I'm planning to discuss the ongoing Trade Directories work at a digital meeting, originally planned for next Monday, but am now waiting to hear when this might be rescheduled. It's a fast-moving area of research and I'm really hopeful we can make some progress towards a usable dataset with structured content; which would be an invaluable project output if we can sketch a roadmap for getting there.
• 8 Feb 2023

  • 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:

    1. 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.

    2. 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.