1648 graphics processed

[vivdunstan]

Happy to be sitting up doing some more old document reading. And thank goodness again for easy colour/balance graphics processing software to make it more readable! Here a testament testamentar (Scottish will and inventory) from 1648. The vast bulk of this 3-page document is a very detailed will. By my 11xg-uncle. Which mentions so many relatives, only some of whom I've figured out. This is related to the academic journal paper I've mentioned recently about the royal secretary (his granddad).

Comments

[a_cubed]

That's going to be a challenge for OCR software!!!

[vivdunstan]

This is vastly better written than some others from the period and earlier I'm trying to read. It's in (mostly) "Secretary Hand", and is nice and clear once you adjust the colours. But yes, transcription software isn't trained on this. I did have fun experimenting with Transkribus (https://transkribus.ai) on 18th century and 19th century text. Unsurprisingly the latter worked better! Though I know some Jacobite scholars who have been using it with some success on mid and late 18th century documents.

Handwriting styles can be learned, but even that isn't enough if the document is in a horrendous state. Here's a 1656 document concerning the uncle of the 1648 will man. And the bleed through is horrific. Yes this is the front side of it.

[a_cubed]

Rather you than me. I have enough difficulty with the writing style of some philosophers (and the poor English of some of the international researchers) without it being physically hard to read, too.

[vivdunstan]

Luckily I like the challenge. It's always satisfying to figure out what something that initially looks like gobbledygook says. And it's an essential part of being a historian focusing on the earlier periods. Only modern historians can dodge handwriting issues a lot.