In the last post, I said that I was planning to change the data structures and page layouts for battles and sieges. I’ve now done that, which means you can:

• view a map of all battles and sieges that have pages on the wiki and coordinates entered on their pages.

[Edit 15 August 2019: the form for searching by date has been temporarily removed because it didn’t work properly and it now needs updating to include other types of event. Most of this post is now obsolete because I’ve merged combat events into a new form/template that can also cover other types of event.]

[Edit 20 September 2019: the search events by date form is back and working properly. It now includes an option to limit the search to certain types of event, as well as searching all events.]

Below is more information about the changes I’ve made.

First of all, I’ve merged battles and sieges into one entity type called ‘Combat event‘. Treating them as separate types was more complicated and confusing. I realised that at a certain level of abstraction, they can be treated as the same thing, and that at a more detailed level, there are two aspects that need to be treated differently in some cases but don’t always go together:

• did it take place on one day or more than one day? This only really affects how data is entered and displayed. Behind the scenes, all combat events have separate start and end dates stored in the semantic data even if the start date is the same as the end date, which means that it’s now much easier to query by date. Single date versus multiple dates doesn’t map easily onto battles versus sieges: a battle or raid could be spread over more than one day, but a town could be captured in one day.
• is it the capture, or attempted capture, of a building or settlement? If it is, then it can have a clear result that is useful to store as semantic data: the place was either captured or held. Results of battles in the open are not represented because they’re much harder to classify and there’s no consensus about the results of some battles, such as Edgehill or both battles of Newbury. This also affects how the sides are classified.

Sides are no longer stored as semantic data as this required five different properties which didn’t have a strong use case. Although they’re no longer semantic data that can be queried, they are still stored as structured data and displayed in a table for the benefit of human readers. If the event is a capture or attempted capture that can have a result, then the sides are classified as besieged, besieging, and relieving. In all other cases, there are only two sides, arbitrarily known as ‘one side’ and ‘other side’.

The table that shows sides no longer shows field signs and words. These can still be added as free text in the ‘Any other information’ section of a page. In future there may (or may not) be a case for reinstating them as structured data, but for now I’m trying to simplify things as far as possible.

Combat events now have a semantic property for WGS84 coordinates. This is a single point somewhere on the battlefield or the coordinates of the building or settlement under attack. I intend this to be used in every case even if the exact location of a battle is unknown, in which case the coordinates can be an approximate location, or the location of a place name associated with the battle. The coordinates are mainly used to place them on the map of all combat events, and can also be used as a fallback if locating them in relation to buildings and settlements is impossible. I still prefer the latter for the maps embedded in each event’s page as it give an impression of the extent of a battlefield without making exact claims about its boundaries, but if no relative locations have been entered, the map will show the event’s own coordinates instead.

Finally, the property ‘Has parent‘ can now be used for any combat event to link it to another event that it was part of (the old property ‘Has related siege’ has been deleted because it was too inflexible). This makes it much easier to represent the battles of Lostwithiel, as well as representing sieges in more detail, as an assault on an individual fortification on one day of a siege can be an entity in its own right with its own result, and can be linked to the siege that it was part of.

Once these changes were made, I had to reimport the existing pages for battles and sieges, which was an opportunity to add more information to them. Pages for battles now have links to the Battlefields Trust website where applicable.

Eventually I’ll fill in more battles and sieges (currently there’s just a selection that was easy to get from Wikidata). I would hesitate to say that all of the data structures and page layouts are now finalised because I need to test some of the other entity types at a bigger scale, but I’m getting there. I hope I’ll be able to make a lot more progress in August because I don’t have too much else to do.