Tag Archive for "uk genealogy" - Chasing Rainbows

Posts Tagged ‘uk genealogy’

Genealogy art

Posted on: June 13th, 2010 by pj

A few years ago I started experimenting with how to present genealogy in something other than a family tree and I combined ‘altered art’ – using an old book, with digital scrapbooking and artist’s trading cards to come up with this:

I found an old book about gardening which seemed the right size and removed some of the pages. I glued the rest of the pages together and to the back board, I then cut out a niche.
I covered the book with mulberry paper, green on the exterior and cream on the interior. I glued aged lace, a photo in a found frame and a buckle. The buckle serves to keep the book closed and to keep the cards in.

I covered the niche and containing page with black lace fabric leaving it to overlap and covered the niche with burgundy crinkled satin-type fabric. I glued a matching ribbon to one side to make it easier to extract the cards.

I glued scraps of tissue newspaper, mulberry paper, a torn photo and a skeleton leaf to the inside front cover.

I made the ATC cards using digital scrapbooking techniques and a mini-book set from fishscraps.com

News from the National Archives

May 17, 2019

On the Record: Double agents and double standards

Espionage is a complicated business. Not only do you, as a spy, have to maintain an alias and feed secrets to your handler, but there is always the possibility that your colleagues or even you yourself could be recruited by another country to become a double agent.

You would then essentially be living three – or more – lives. The first being the one that the outside world sees; the second as a spy for your country; and the third – the most secret of them all – as an agent for another state.

This is what happened to members of the Cambridge Five spy network, who became the focus of probably the most well-known real life spy story in British history. For years their stories remained hidden, and in many ways masked by popular speculation, but with the help of archives, historians, and records experts it is possible now to uncover the real story behind what we think we know.

This is what The National Archives is aiming to do with our new podcast series, On The Record. The first topic we are tackling as a mini-series is espionage; using previously secret personnel files and government reports we uncover the true stories of famous spies. In episode one we focused on the ‘Archetype of a Spy’, and episode two looked at the real stories of T. E. Lawrence and Gertrude Bell, who worked in the Middle East gathering intelligence and influencing politics.

In the third and final episode of the series we examine the stories of individuals accused of spying for foreign countries, including those who were part of the Cambridge spy network. I spoke to Stephen Twigge about the Cambridge Five and why they are so well-known as double agents.

Photographs of Maclean and Burgess from post-disappearance papers 1951-1953. FCO 158/6

[Katie]: Why do you think the story of the Cambridge Five has become so well-known?

[Stephen]: It’s probably the enormity of the betrayal and the fact that they all came from privileged backgrounds. Communists were supposed to be working class, not part of the establishment studying at Britain’s top university. The volume of books and films on the subject has also helped to develop the mythology of the Cambridge spy ring.

Draft minute from March 1952 regarding the Burgess and Maclean case. FCO 158/27

[Katie]: Do you get any sense from the documents of how the men were able to manage the various pressures of being double agents?

[Stephen]: At first they seemed to manage well, but the pressure of maintaining a double life eventually became too much and they began to crack. Files released recently show that their Soviet handlers viewed them as hopeless drunks incapable of keeping secrets.

[Katie]: There are other people who have been named as additional members of the spy network. Do you think this speculation has helped to popularise the story and make it as well-known as it is today?

[Stephen]: Yes, there is a small cottage industry producing books on the subject and every now and again a new name emerges as having been part of the ring. Those recently ‘unmasked’ include Victor Rothschild, Guy Liddell and Roger Hollis. Whether there is any substance to the allegations is debatable but that doesn’t stop people speculating and adds additional layers of intrigue. It will probably keep people guessing for years.

December 1951 letter from the Foreign Office to Sir Oliver Franks, UK Ambassador in Washington, advising Franks to burn this and the previous letter. FCO 158/27

You can find out more about the Cambridge spy ring in a previous blog on the subject.

On the Record: a new podcast series

All three episodes of On the Record are available on the Archives Media Player as well as on podcast listening apps.

Subscribe: iTunes | Spotify | RadioPublic | Google Podcasts

To find out more about espionage, as well as life in Britain during the turbulent Cold War era, you may be interested in The National Archives’ Cold War season, which includes a new exhibition: Protect and Survive: Britain’s Cold War Revealed.

The post On the Record: Double agents and double standards appeared first on The National Archives blog.

May 13, 2019

How to teach a computer to read

A thought that may have crossed your mind when reading the title of this blog is ‘Why would we want a computer to read in the first place?’ The answer to this is that, despite the overwhelming size of The National Archives’ Discovery catalogue, there is so much more information out there! Computers can help us reveal this previously hidden information.

Take, for example, PROB 11, a vast set of registered copy wills from the Prerogative Court of Canterbury. These range in date from 1384 to 1858 and contain a wealth of information about places, beneficiaries, family, relationships, goods, religion, values and land, to name but a few. Whilst one will can contain information on multiple people, the catalogue description only shows the will maker, place and a date, making the rest tricky to search. With 2,000 volumes each containing thousands of pages, adding the extra information would be an enormous task for volunteers, to either extract the more complex data or transcribe the documents. It is in these scenarios that we could really use some help.

An example of a will from the PROB 11 collection compared to its catalogue description

Recently I completed an Early Career Fellowship at The National Archives, the first to be very generously sponsored by the Friends of The National Archives. The position was part of the Digital Research Team, helping with one of their current research projects, namely the transcription of PROB 11 using machine learning. The topic was suitably broad, looking at ways of crowdsourcing in conjunction with the processes of the handwritten text recognition software, Transkribus.

Transkribus is software that anyone can download, which can read handwritten documents. To do so, it requires full transcriptions of a sample of documents. The sample documents and transcriptions are given to the software which then creates a model of the handwriting, which it can use to read further examples of this specific style of handwriting. Once results start being generated from the model we can improve the accuracy by correcting it, producing better and better results.

The Crowdsourcing Cycle

When it comes to learning, computers are very similar to humans: more examples and more experience make them much better at reading. If you ask a computer to read a style of handwriting it hasn’t seen before, it probably can’t, and words that it sees less frequently are less likely to be correct. This is why collections such as PROB 11 are perfect to trial with this type of software, as they are very formulaic and the handwriting style is consistent.

But how can we help? There are two points of human input in the whole process – the full transcriptions at the very beginning and at the correction stage. A lot of people enjoy transcribing documents but correction is seen as a more mundane task, which is not as much fun. However, from an efficiency perspective it takes much less time to correct documents. A large part of this research project was about making correction more enjoyable and easier for the user, looking at their key motivations to transcribe documents. This research can be divided into three key areas.

Table showing user motivations to transcribe documents

The first key area was to find a way of visualising correction on a large screen. To isolate areas most likely to be incorrect and make them easier to find, we used Natural Language Processing to generate probabilities of how likely Transkribus was to be wrong. Then using these probabilities we highlighted particular lines that needed the most correction so that users could specifically target them. A lot of fun was had using different colours for the highlighting.

When looking at motivations to transcribe documents, volunteers often enjoy knowing the context and history behind the work they are doing. By displaying a full page there is the opportunity to not only correct but also to read the full document.

The second area was based on an overview look at crowdsourcing projects across multiple sectors. The most exciting projects had two or more volunteer motivations behind them. For example, a project where a group go for a run and also paint a community centre, both helping the community but also encouraging fitness and the social life of the participants. Volunteering through which the participants learnt or developed skills was also highly successful.

Another big factor was time-saving initiatives. If you can do two or more things at the same time then it was more likely to be successful. Crowdsourcing and volunteering should fit in to people’s lives and the rise of micro-volunteering initiatives demonstrates that.

The average UK citizen spends around about an hour commuting each day. As a result we did some research into the idea of apps and transcription on the go with the additional purpose of being able to teach the user palaeography at the same time. We created a demo version of an app called ‘Many Hands’, which gives the user exercises to improve their palaeography skills on the move whilst also correcting transcriptions from Transkribus.

Documents from the PROB 11 collection highlighted to demonstrate potential inaccuracies

Finally, we looked closely at ways to make transcription fun for those who might not know about it at all. As a result we ran a Game Jam with the Transcribe Bentham Project at University College London (UCL) and the UCL Innovation and Enterprise team earlier this year. Further details can be found here and here in previous blog posts. The aim of the Game Jam was to make the process of correcting and transcribing more fun and efficient. Around 55 people attended with a variety of interests and perspectives, ranging from academics to game designers, and made some amazingly inventive submissions. The winning submission was a version of the game Frogger, where the player has to dodge and transcribe criminal records from HO13. See the results for yourself! It was a great success and we came away not only with lots of fantastic new ideas, but with confidence that transcription can be fun for everyone!

Our game jam entries

A big thank you to the Friends of The National Archives for supporting this research, and the staff at UCL and the Transcribe Bentham Project, specifically Dr Louise Seaward, who made it possible for the game jam to take place.

The post How to teach a computer to read appeared first on The National Archives blog.