The British Library situation

If you have anything to do with historical research, you probably know that most of the British Library’s online services have been out of action since October 2023 because of a ransomware attack. The knock-on effect for By The Sword Linked is that outgoing links to UINs in the Main Catalogue, Archives and Manuscripts Catalogue record permalinks, ESTC numbers, and EThOS thesis IDs are all broken. This incident shows a problem with linked data: if one website goes down it also affects other sites. But it also shows that making Open Access publications and reusable datasets available in more than one place can limit the damage.

The good news is that the Shared Research Repository doesn’t seem to have been affected. The repository makes available Open Access books, reports, and datasets. This includes all the thesis metadata from EThOS, last updated in November 2023. In 2021, I used an earlier version of this dataset to add wiki pages for 600 theses relevant to the British Civil Wars, linked to pages for authors and subjects (and many of the authors are linked to Wikidata IDs). In many cases, the EThOS data includes a link to the full text of a thesis at an institutional repository.

This week I’ve imported wiki pages for some British Library manuscript collections and volumes. I compiled the data from my own research notes and the printed catalogues at the Internet Archive. Where I’ve found printed catalogue entries, the wiki page for the collection links to the Internet Archive (although you might find the links unreliable because the IA servers are overloaded). Some collections are linked to subjects and creators, but many aren’t because they’re too miscellaneous. The easiest way to see what’s there is to drill down from the page for Western Manuscripts. I’m working on more data for British Library manuscripts, but I don’t know when I’ll have time to finish it because I’m starting a new job soon. Meanwhile, I’ve put up a general list of interim catalogues and finding aids.

New content imported

Before and after the server move, I imported some new content to the wiki (at last!):

  • Most peers and peerages of England, Scotland and Ireland, 1630-1669. There are probably errors and omissions because I mostly imported the data from Wikipedia. Many subsidiary and courtesy titles are certainly missing. As well as drilling down from the links, you can find individual holders of titles by typing in the search box. The page names start with the person’s surname, and there are redirects that start with the distinctive part of the title. For example, Essex, Earl of, 3rd (Robert Devereux) redirects to Devereux, Robert (3rd Earl of Essex). Arranging this data led to creating a new property: Ordinal. This is the number of a title holder such as a peer or baronet.
  • MPs in the Long Parliament now have constituency links, referenced to Brunton and Pennington. There may be some mistaken identities where I had to rely on Wikipedia for disambiguation, and a few MPs are still missing because I couldn’t positively identify them. Things will improve once Andrew Gray has imported History of Parliament data to Wikidata. I’ve temporarily removed links to the Short Parliament because the data I had was so inadequate and it will be easier to redo it properly with nothing there.
  • Every surviving Buckinghamshire loss account that I know of now has a page and is linked to the place it’s about, so they should show up in the query for linked sources on pages for the relevant settlement.
  • All (probably) relevant articles from Journal of the Society for Army Historical Research. Most of these link to archived copies at JSTOR, except for the last two years, which are still behind the moving wall.
  • The Making of the English Landscape series now has all the county volumes that were published (the series was never finished, so some counties are missing). These link to National Character Areas covered as well as counties.
  • Contents of Midland History and Helion’s Century of the Soldier series have been updated to the end of 2023.

Maintenance finished

The wiki is now up again. The address is the same but it’s hosted on a different server with newer versions of Mediawiki and extensions. This has fixed some minor problems, although there could be some new problems that I haven’t found yet. Before and during the upgrade, I made some changes to the data structures:

  • Has URL is now used for units. This is their official website if they have one now (such as London livery companies). For regiments, it’s a link to the BCW Project Regimental Wiki, although that site is temporarily down.
  • Units have new properties: Has clothing and Has symbol. These are mainly intended to represent coat colours and flags of military units, but they can also be used for coats of arms, livery etc.
  • Removed external identifier properties for people, except Wikidata ID, which is now the spine for all other person identifiers.
  • Book editions have been split into separate forms and templates:
    • Early edition. Published before 1800. Corresponds exactly to an ESTC number and has no lower levels.
    • Modern edition. Published after 1800. Formats are now represented by repeatable subtemplates within the edition page, and never by separate pages.
  • Work level for books and articles is now only represented when needed: when a work has more than one version, or a Wikidata ID.
  • Printed copies and printed text sections are no longer represented. They could be brought back in future, but for now the main focus of the project will be more on manuscripts than print.
  • Property for author death year removed. This can usually be discovered via Wikidata.
  • Removed property for gender status, but personnel roles are now classed by gender, which is a more efficient and flexible way of classifying people by gender.
  • New property Is role in, used to link a personnel role definition to the types of organization that it can be used with.
  • Manuscript texts no longer have properties for date signed or the date an original will was made. These are now covered by Has earliest known date.
  • Ancestor command relationship now has more specific subclasses, which make it easier to query for things like all the settlements in a county.

Downtime for Maintenance

The By the Sword Linked wiki will be temporarily unavailable because of a server move and software upgrade. I can’t predict exactly when it will go down and come back up but it should all be done within 2 weeks. I’ll post again to say when it’s back and what has changed. If anything goes so badly wrong that I can’t bring it back, the entire contents will still be available at Github under CC-BY-SA.